Mobile Innovations Forecast

Contents

1. The elements of contextualintelligence

2. The contextual network

3. Device and environmentunderpin contextually awareservices

4. New technological capabilities

5. New data bolsters the generaldirection of innovation over thenext five years

6. Enabling devices to offer usersmore natural interaction

7. Image sensor: Steady growthfor new capabilities

8. Storage: Quenching the thirst

for more

9. Memory: The ever-predictableDRAM path

10.Application processors: Drivingthe next wave of innovation

11.Infrastructure speed: Watchcapital investment in 4G for thenext inflection

12.Device connectivity speed: Onehalf of an equation

13.Making sense of the rapidchange in mobile innovation(includes PwC's MobileTechnologies Index)

14.Mobile operating system:Smartphones will just getsmarter

www.pwc.com/technology

Mobile Innovations Forecast The elements of contextual intelligence

Today’s mobile users want to navigate their daily lives using a combination of highly personalised mobile devices, information and applications. Mobile operators covet the revenue these users will generate. Brands know they need to be prominent in this mobile bundle if they want to engage this audience. And no serious enterprise discounts the role mobile will play in its future.

However, the mobile user experience is inspired more by desktop metaphors than by the always present, highly personalised reality of mobile devices. Mobile interactions too often involve ‘mobilised’ desktop applications or websites shrunk to fit the mobile screen. Even native mobile apps have limited awareness of what a user wants without explicit input.

Technology Institute

Synopsis

This is the fourth article in the Mobile Innovations Forecast Phase II: New technological capabilities. Here is a roadmap of the series:

- The Introduction argues that the dominant drivers of mobile innovation to 2018 will revolve around capturing and modeling the contextual situation of mobile users, and will transform the mobile device into an intelligent digital assistant.

- The second article examines how device and environmental sensors interact to capture information to model the user’s physical context.

- The third article examines how communications networks enable interaction of the user’s physical context data with information and applications in the cloud.

- This fourth article describes the capabilities that enable the mobile device to generate contextually relevant information and services.

- A concluding article will highlight the most significant new capabilities driving smart devices towards true digital companionship, setting the stage for new use cases and business models to follow.

Raman Chitkara, Global Technology Industry Leader

But the ground is shifting. As mobile markets mature, value creation is migrating to how well a device or service adapts its content and functionality to a user’s needs and preferences. Having awareness of the user’s contextual situation—where she is; what she likes; who she knows; how she has previously used a device or service and so forth—and having the intelligence to act on that knowledge is becoming a core driver of mobile innovation. As this process iterates, contextual awareness and intelligence will become a major source of growth for both the mobile and information technology industries.

http://www.pwc.com/technology

http://www.pwc.com/gx/en/technology/mobile-innovation/technology-predicts-real-time-behaviour.jhtml



http://www.pwc.com/gx/en/technology/mobile-innovation/assets/phase2-advancing-technology.pdf

http://www.pwc.com/gx/en/technology/mobile-innovation/assets/pwc-mobile-innovations-forecast-the-device-environment.pdf

http://www.pwc.com/gx/en/technology/mobile-innovation/assets/pwc-mobile-innovations-forecast-the-contextual-network.pdf

Mobile Innovations Forecast: Phase II / 2

This article examines the emerging capabilities that enable mobile devices and services to provide contextually intelligent information or take actions on behalf of the user in a natural, almost humanlike, fashion. PwC believes that contextual intelligence arises from coordination across three domains of technologies and capabilities found in the mobile device, the telecom network and the cloud:

a) Conversational intelligence: The natural language interfaces that converse with end-users on their mobile devices and deliver most contextual experiences to them;

b) Sensor intelligence: The software development environments that add contextual capabilities to mobile applications and services;

c) Decision intelligence: The machine learning capabilities that derive personal behaviour patterns to predict when and how to proactively engage the user with information, recommendations, and actions.

These domains interact at multiple levels and across multiple players. No single entity or industry is likely to own and control the end-to-end contextual value chain. Rather, the success of contextual technologies and services is likely to be ecosystem-led more than technology feature-led. With that in mind, it’s time to explore each domain.

PwC defines contextual awareness and intelligence as a form of computing in which situational information about people, places and things augments the more slowly changing personal profiles of users and is leveraged to anticipate an end-user’s immediate needs. Based on that knowledge, contextually intelligent services proactively offer enriched, situation-aware and actionable content, functions and user experiences. More importantly, every interaction adds to the depth and breadth of insight about a mobile user.

Since 2011, the most well known example of a contextually intelligent mobile service is Apple’s Siri, a voice-activated virtual assistant that responds with

Contextual awareness and intelligence defined

user is more often navigating locations, choices and relationships that change dynamically as she progresses through her day.

The recurring question for mobile users as they encounter new information, people or choices seems to boil down to ‘what is the best next step for me to take right now or in the near future?’ The ability of a mobile device and service to be ready with contextually relevant information and services helps people decide what those next important steps should be.

information and services that grow in relevance the more the system is used. Other general-purpose virtual assistants such as Google Now or Microsoft’s Cortana have launched to establish positions in the nascent market for mobile contextual services.

Whether embodied in a virtual assistant or integrated into a specific application, contextual awareness and intelligence is especially valuable for mobile user experiences. Unlike when she uses a desktop computer, the mobile user typically is not navigating documents or websites to meet her needs. The mobile


Conversational intelligenceContextually aware and intelligent experiences require users to educate their mobile devices or services about their needs, preferences and desires. Some of this education will be explicit, such as a user giving a voice command to his device. Much more of this process will happen in the background.

Whether users actively or passively exchange information with their devices and services, their contextual interactions, whilst mobile, are likely to be conversational rather than based on formal commands or menus. “Siri, tell me which exit I should take to go to Northgate Shopping Mall” is a far more intuitive—not to mention safer—user experience than trying to input key words into a search box, and then read the results whilst driving a car. If the user’s conversational experience is his initial exposure to contextual services, then the quality of the conversation significantly influences his decision to adopt contextual services in the first place.

Consequently, a core capability for contextually aware and intelligent mobile services is natural language processing (NLP). NLP refers to the ability of a computer to understand and converse in human language as human beings communicate with it. NLP is a hybrid discipline that combines expertise in computer science, artificial intelligence (AI) and linguistics. It is the technology and information foundation for the natural language interfaces used

by virtual assistants like Apple’s Siri, Google Now and Microsoft’s Cortana. As such, NLP has a profound impact on the contextual experience for the vast majority of users.

Writing in the Chronicle of Higher Education1 , Dr. Geoffrey Pullum, a professor of general linguistics at the University of Edinburgh, said that effective NLP results from the mastery of three core tasks by computer systems. Assuming that speech recognition technologies have extracted human words and phrases from the ambient environment, an NLP system must first understand syntax of a target human language to uniquely identify one sentence or phrase from another sentence or phrase. Second, an NLP system must possess semantic capabilities to extract the literal translation from a sentence or phrase and relate it to translation of other sentences and phrases. Finally, and most difficult, an NLP system must possess the pragmatic rules of thumb to infer the intent behind a human utterance, and thus discern what should be assumed or performed given the meaning of a sentence or group of sentences.

Although NLP has been a recognised subcategory of AI research since the 1950s, it was not until 2011 that NLP became well known to the general public. NLP made its most recent public debut in February 2011 when IBM’s Watson supercomputer defeated two

1 http://chronicle.com/blogs/linguafran-ca/2013/05/09/natural-language-processing/

human champions at Jeopardy! While a flurry of positive and negative stories tried to untangle the ultimate impact of a computer’s victory over humans in a contest conducted in natural language almost all observers were impressed by Watson’s ability to interact smoothly with its human competitors and the game show host.

Then, in October 2011, NLP took another mass-market step with the launch of Siri on the iPhone 4S. Siri was the first mass market, general purpose2 virtual assistant that employed NLP to create an interface in which users could speak in free-form natural language to query for factual output (“what was the score of last night’s game between X and Y teams?”) or perform voice-activated functions (“Siri, please send a message to the person I’m meeting at 3 p.m. that I will be 20 minutes late”). More recent virtual assistants, such as Google Now and Microsoft’s Cortana, combine spoken and text-based interfaces based on NLP to expand the range of queries and tasks that can be executed by the user.

PwC believes that broad adoption of contextual intelligence pivots on the ability of users to engage with their devices and services in a conversational manner rather than in a command and menu approach. Just as the graphical user interface transformed the desktop computing experience for ordinary people, and hyperlinking defined how people experienced the World Wide Web, NLP enables a fundamentally new interface amongst people, information and technology.

These activities happen continuously at multiple levels of the mobile device, the network and stored data and/or functionality in the cloud. Depending on the nature of the contextual experience, these activities can execute across a number of service providers and ecosystems.

2 Domain specific NLP systems for applications such as stock trading have existed for over a decade; Siri was the first to successfully remove the assumption of a specific seman-tic domain.

Element Capability

Syntax Uniquely identify one phase/sentence from another

Semantic Extract literal translation

Relate to other sentences/phrases

Pragmatic Apply rules of thumb to phrase/sentence to discern what should be assumed or performed given the meaning of the phrase/sentence

Source: PwC

Table 1: The elements of natural language processing


Contextually intelligent systems interact and learn about humans in real-time by incorporating the following information and processing:

1. Acquiring information about a user and his or her environment. The system typically draws its raw contextual data stream from physical sensors in the device or immediate surroundings, plus stored information on the device or in the cloud. Observation of contextual data might be episodic and initiated by the user, but most often is passive and continuous.

2. Modeling the current circumstance or intention of the user. A contextual system usually operates according

Common attributes of contextual systems

to communicate with the user. The system also communicates with data sources and applications through APIs or similar means.

5. Learning through feedback. A contextual system continually analyses the conditions that trigger an event; the information and applications used to generate a contextualised response; plus the user’s positive or negative reaction to the output offered by the system. These results feed back into the contextual system’s configuration to provide increasingly personalised experiences to the user the more the system is used.

to pre-programmed heuristic rules to build a predictive model of user intent. These rules are refined and augmented as the user repeatedly accesses contextual services.

3. Reasoning about the next best step and taking action. Based on its predictive model of user intent, a contextual system can access applications and data to offer and deliver contextually relevant experiences. The back-end resources that enable these experiences may or may not be known to the user.

4. Notifying a user or other apps/services of an action. Contextual systems use conversational intelligence, based on NLP capabilities,

1

2

34

5

Based on the user's reaction, the system provides information or functionality, logs trigger conditions and data for the contextual database, and adjusts the predictive intent model another iteration.

A predictive model of user intent ingests and processes contextual information

The model generates best next steps for the user

A contextual database accumulates sensor data plus stored user information

The device engages the user in conversation.

Contextually intelligent systems ‘learn’ in real time about humans and their environments through the exchange of information.

Source: PwC


Sensor intelligenceIf NLP provides a conversational interface to contextual experiences, then contextually intelligent applications use voice and sensor data to turn those conversations into action. This section explores the four primary contextual data inputs that help make a mobile application contextually aware and intelligent. It then provides a recent example of a context-focused software development kit (SDK) for mobile retail that packages these inputs so non-specialist developers can create contextually informed applications.

The four primary contextual data inputs (location, identity, time and activity) not only answer questions of who, what, when and where; they provide the contextual data resource for local applications

and decision engines in the cloud. For example, a contextual application might access location and time information from a context SDK operating on a mobile device. If the location data of a user in her home corresponds to the kitchen, and if the time stamp is 0700 local time, the statistical correlation score for the contextual situation of ‘breakfast’ is suitably enhanced. Based on that insight, the contextual application will return some type of information or take an action according to the design of its developer.

Thus, it is useful to think of location, identity, time and activity as the primary colors of a contextual palette. Individually, these data types provide contextual value to an app depending on their presentation and use. More important, these inputs can be combined or mixed (location + activity) with secondary inputs, such as personal history, to create a range of options for adding further contextual value.

One of the first SDKs designed for building mass-market contextual applications is Gimbal from Qualcomm Retail Solutions. Gimbal uses location plus personalisation features to add contextual value to mobile apps focused around the experience of physical retail.

The platform is built around three core concepts:

1. Location—Where are the customers?

2. Proximity—What are the customers near?

3. Interest—Who are the customers and what matters to them?

In a typical scenario, a user downloads a Gimbal-enabled mobile app to share her location and other information on an opt-in basis in exchange for something in return. When the user’s mobile device crosses a geofenced3 boundary, that event triggers an action on the device or in the cloud that returns some type of

3 A geofence is a virtual perimeter around a real geographic location. Geofencing applications and services allow an administrator to set up triggers so when a mobile device enters (or exits) the boundaries defined by the administra-tor, a message is sent or a processing event occurs. Source: http://whatis.techtarget.com/definition/geofencing

Figure 1: The four primary contextual data inputs

Location

Identity

Time

Activity

Source: PwC

Integration of primary and secondary data inputs creates contextual intelligence.


information or digital token to the user. The output may be a standard offer for each opt-in participant or the output may be customised based on stored individual preferences or history. Simultaneously, the Gimbal platform records activity data for follow-on analysis by an application developer, a retailer or venue owner.

In addition to enabling application developers to draw macro-level geofences around physical locations such as a store or a shopping mall, Gimbal offers micro-location capabilities for areas 50 metres or less in radius by using proximity beacons4. These in-store devices are based on Bluetooth Low Energy (LE) to transmit ID codes that can be picked up by mobile devices that have downloaded retail applications. When a device is within physical proximity to the beacon and detects it, an app can notify the customer of location-relevant and individual-relevant content, promotions and offers. On the back-end, the app developer can set various interactive rules for the beacons such as activating during specific hours or engaging only with a certain shopper profile when she is within proximity.

The ability for non-specialist application developers to use an SDK like Gimbal to merge contextual inputs such as time and identity with fine-grained location capabilities for a target user experience is a major advance for contextual services. As demonstrated in other markets, innovations do not grow to mass-market status by staying rooted in the industry sector that spawned them. For contextual experiences to scale, developers that specialise in particular areas, such as retail, must be able to integrate sophisticated contextual capabilities into their applications without being contextual technology specialists. In that sense, general purpose contextual SDKs are a significant ingredient for success.

4 A proximity beacon marks a radio-based enclo-sure that a developer sets up around a smaller geographic area; for example, a department within a store, a specific street address, a sec-tion of a parking lot or a landmark in a plaza. Source: www.gimbal.com In addition to Gimbal, Estimote, Kontakt, and GeLo, to name a few, are offering proximity beacons.

Decision intelligenceIf the purpose of natural language interfaces and contextual applications is to package and present a contextual experience for a human or piece of mobile technology, then machine-learning capabilities create the ‘decision intelligence’ that make the experience possible.

Machine learning represents a significant departure from traditional system development methodologies. For most of computing history, programmes were built by distilling knowledge from human experts into a series of logical structures that enabled a system to respond in predictable, repeatable ways. If you wanted to build an accounting system, you started by interviewing human accountants to understand and create the rules that software engineers encoded into formal logic that could be understood by machines. So long as a target process lent itself to high levels of formalisation, the methodology worked reasonably well.

However, highly formal systems don’t handle ambiguity or exceptions very well. Take NLP. Humans have tried and failed numerous times to develop a complete but manageable set of formal language rules that can handle the standard tasks and the exceptions of translation. Not only is human language rife with exceptions due to regional dialects and a host of other idiosyncratic factors, it is constantly evolving.

But by building a framework that enables software to start with some pre-programmed examples of previous, successful translations and then to compare those examples with a new sequence of words, a computer system might get closer to making a successful new translation. Add in scoring mechanisms for the system to track whether its current translation is closer or further from a target accuracy, and the system gains the ability to adjust its processing for the next translation attempt. Over time, the system will ‘learn’

Machine-learning capabilities create the ‘decision intelligence’ that make the experience possible.


Figure 2: Machine learning provides structure to unstructured information with minimal human involvement

to recognise statistically significant translation patterns that should grow in accuracy the more the system is used. In plain English, a machine learning system distills the rules it requires from the data on which it is exposed and trained, rather than having all that knowledge directly coded by the programmer.

Machine learning is foundational to contextual systems because it offers the ability to sift through vast data sets and classify preliminary patterns in a user’s contextual data stream without direct human intervention. Sensor data logs, user transactions, check-ins, captured media, repeated location visits—all of these and many more will be sifted for patterns that fine tune predictive algorithms that anticipate, engage and perform actions for humans.

Based on these patterns, a key output of machine learning engines will be to place a human user into a contextual knowledge graph. This graph combines literal intelligence about the user’s documented habits and emotional inferences about his typical states-of-mind (example, no alerts or interactions during sleeping hours). A knowledge graph is further enhanced through mapping relationships, classifications and genres derived from the four primary contextual inputs described above. This structured information is then made available to contextual applications for eventual presentation to the user.

The ability of machine learning to provide structure to unstructured information with minimal human involvement lies at the heart of its value proposition. For example, today’s virtual assistants are highly responsive and accurate in providing information about movies playing at local theatres. The movie is playing at the theatre or it is not. However, analysing previous user interactions around movie content plus the preferences of her social network and her current location and time to make a recommendation about a new movie she might like, requires a level of processing only machine learning can provide.

Location ID Time Activity

Sensors and environmental beacons generate primary level contextual data points.

Primary level contextual data is sifted, correlated and augmented with secondary data drawn from user history and other contextual inputs.

Machine learning engines derive patterns in user behaviour and intent to place her on a contextual knowledge graph.


smartphones to include a range of interactive situations (e.g., Amazon FireTV, connected cars, e-commerce, tele-medicine). Marketers and app developers will need to prepare for this fundamentally different model of customer relationship management.

2) Contextual SDKs like Gimbal will open up contextual app development to a much larger base of developers many of whom will create industry-specific contextual applications. Some early examples might be contextual apps for healthcare, education, travel, fitness and wellness. Technology companies with seemingly secure roles in these domains should be building out mobile contextual apps before new entrants challenge their market positions.

3) Machine-learning systems will encompass a range of contextual processing tasks in addition to NLP. Most likely, these systems will focus on domain-specific contextual knowledge to support industry-focused contextual SDKs. Over time, more individual domains will connect and integrate with one another to evolve general-purpose contextual assistants to more sophisticated and personalised levels. Companies with little or no machine-learning experiences should be exploring ways they can augment user experiences beyond the current standard of certainty (i.e., what is the purchase order number), and into areas of uncertainty (how reliable will this new supplier be).

The net result of these trends will be a veritable explosion of the contextual information market ecosystem, far beyond what is seen today. Activities, locations, transactions, preferences, emotional

states—all of these will be logged, compared for patterns and archived to create the contextual experiences that drive JIT mobile markets.

Managing privacy will become one of the biggest and most important industry sectors for the contextual age. So too will be the industry players dedicated to managing and visualising the contextual information spaces of users.

Daniel Eckert

Managing Director, Emerging TechnologiesPwC

“There are significant opportunities for certain players with heavy infrastructure and analytic resources to catalyse entire ecosystems around contextual data,” says Daniel Eckert, Managing Director of Emerging Technologies at PwC. “For example, data about weather is free from the government—but once the data is enriched by a contextual data provider for a particular audience—it can be utilised for many different services.”

The opportunities are immense for contextual services as are the challenges, both technical and social. But the inherent value of enabling users to converse naturally with their devices and service to get what they want, when and how they want it is hard to deny.

Machine learning in 2014 is focused mainly on improving NLP accuracy and performance. But the underlying principles of machine learning are applicable across the contextual value chain. Equally important, machine learning is the means through which contextual capabilities scale to the mass-market. Humans repeatedly prove that they will use any and all enhanced computing capabilities offered and then demand more, analogous to the way devices will use storage, application processing speed and other enabling technologies covered in Phase I of the Mobile Innovations Forecast. It follows that the more contextual intelligence powered by machine learning is added to an interaction, the more sophisticated user behaviour becomes, which creates more opportunities and demand to add value.

Conclusion and forecastContextual intelligence enables a ‘just-in-time’ (JIT) mobile lifestyle that is becoming more prevalent. Cultural studies suggest that mobile users, especially younger demographics, are using their devices as much for organising their daily lives as for communication. Given that reality, it is clear that the days of the generic mobile user experience are numbered. Users will expect experiences specifically tailored for them that evolve in the face of new situations.

PwC believes the impact of contextual intelligence will be broad and deep—affecting every part of the mobile ecosystem. We anticipate three general trends that decision makers should consider as they develop strategies and capabilities for their organisations:

1) Moving beyond its roots in narrowly defined situations like stock trading, NLP transforms human computer interaction. Voice-controlled interfaces enabled by NLP will proliferate beyond

PwC refers to the PwC network and/or one or more of its member firms, each of which is a separate legal entity. Please see www.pwc.com/structure for further details.

This content is for general information purposes only and should not be used as a substitute for consultation with professional advisors.

© 2014 PwC. All rights reserved. PwC refers to the PwC network and/or one or more of its member firms, each of which is a separate legal entity. Please see http://www.pwc.com/structure for further details. CH-13-0114

Let’s talkIf you have any questions about the Mobile Innovations Forecast or would like to discuss any of these topics further, please reach out to us.

Raman Chitkara Global Technology Industry Leader PricewaterhouseCoopers LLP [email protected]

Daniel Eckert Managing Director, Emerging Technologies PricewaterhouseCoopers LLP [email protected]

About PwC’s Technology InstituteThe Technology Institute is PwC’s global research network that studies the business of technology and the technology of business with the purpose of creating thought leadership that offers both fact-based analysis and experience-based perspectives. Technology Institute insights and viewpoints originate from active collaboration between our professionals across the globe and their first-hand experiences working in and with the technology industry. For more information please contact Raman Chitkara, Global Technology Industry Leader.

About PwCPwC helps organisations and individuals create the value they’re looking for. We’re a network of firms in 157 countries with more than 184,000 people who are committed to delivering quality in assurance, tax and advisory services. Tell us what matters to you and find out more by visiting us at http://www.pwc.com


Mobile Innovations Forecast Virtual context: Connecting two worlds

Contextual services will accelerate the shift to programmable networks When people upload mobile photos or video to a social network, back up documents and media files to the cloud or sync their data from a mobile device, they typically cross multiple network boundaries and data centres. In these scenarios, latency, jitter and packet loss affect user perceptions of network quality just as pops, crackles or dropped calls impact perceptions of mobile voice quality.

These data quality metrics have always been important, but they take on greater significance as people use increasingly sophisticated, context-aware mobile devices and services. As Qualcomm CEO Dr. Paul Jacobs has noted, wherever users physically travel, they will be at the centre of a personal cloud of nearby devices, apps, information and contextual


Synopsis

This is the third article in Mobile Innovations Forecast Phase II: New Technological Capabilities. Here’s a roadmap of the series:


- The second article examines how device and environmental sensors interact to capture information to model the user’s physical context.

- This third article explores how communications networks will enable interaction of the user’s physical context data with information and applications in the cloud to create the virtual context layer.

- The fourth article will describe the modeling, intelligence and analytic engines, mainly in the cloud, that will enable the mobile device to become as intimate as a personal assistant, if you allow it.


Raman Chitkara, Global Technology Industry Leader

choices. “We’re working on a vision we call the Digital Sixth Sense, this idea that the world around us will be connected, and the phone and devices we carry will allow us to essentially blur physical space and cyberspace,” he recently told PwC in a interview for their 17th Annual CEO Survey.1 Now imagine a sensor-connected, visual media-centric world in which the end user’s personal cloud of devices, information and applications continuously interacts and exchanges data with numerous service provider clouds depending on various triggers, such as geofences, in the physical world.

1 http://www.pwc.com/us/en/ceo-survey-us/2014/assets/dr-paul-e-jacobs.pdf

Pierre-Alain Sur, Global Communications Industry Leader




This article focuses on information exchange between the digital and physical worlds that creates the virtual context of the end user. It is concerned with how the networking requirements of contextually aware and intelligent services will enable the dynamic environment described by Jacobs. It builds upon previous work in this series that focused on enabling technologies in Phase I,2 and the physical context of users identified in the preceding article.3 A fourth article will examine how cloud-based analytic and predictive engines organise and make sense of the massive data flows handled by mobile networks to create a seamless, contextualised user experience.

A killer ecosystem rather than a killer appThe virtual context of a user is created by the layer of telecommunications technologies and services that connects situational data captured by the mobile device with data, analytics and applications in the cloud. Telemetry from device sensors, location/object beacons, user-generated media and physical motion is uploaded into networks whilst interactive maps, augmented reality visualisations, streaming media and other contextually relevant information is downloaded from networks for display on a user’s mobile device.

Virtual context is a dynamic environment in which a user’s personal cloud of devices, data, preferences, applications and social connections constantly updates and adjusts its behaviour based on its interactions with the physical world plus information and applications accessed from networks. “You want to integrate a specific location and a specific activity you’re doing with the digital information you have from your social network, your email communication,

2 Wrapping up Phase I: New data bolsters the general direction of innovation over the next five years

3 Sensing and making sense: Device and envi-ronment underpin contextually aware services

the notes you take on the device and so forth,” says Oliver Brdiczka, director of contextual intelligence research at PARC, a Xerox company.

Making such an experience robust and seamless to the user requires significant network bandwidth. However, raw capacity alone is not sufficient to enable diverse mass-market contextual services. Networks must become flexible enough to not only provide various levels of quality of service, but also handle security and other back-end services such as billing for end users and third-party service providers. For that to occur, communications networks must become more than just large data pipes. They must enable the digital equivalent of smart logistics for telecom and non-telecom service providers for whom the network is the front door to the customer.

For example, a media company might want the communications network to emphasise low-latency for streaming files with a thin path coming back. At the same time, a gaming company looks for low-latency for both the uplink and downlink. Healthcare service providers need network performance plus security that is technically compliant with regulatory objectives. Search engines will start to receive images uploaded by the user in addition to key words before returning a result.

Practically speaking, the need for flexible, robust digital logistics will push cloud computing principles and technologies deeper into the design and operation of communications networks. Both communications infrastructure and service providers are virtualising telecommunications in much the same fashion as the IT industry virtualised computer processing and storage. This activity is accelerating the creation of a new capability—the programmable network.

The rise of programmable networks Programmable networks are those in which software directs the flow of data and the behaviour of network elements in a manner that is largely independent of physical hardware. It is a process not dissimilar from the unbundling of software from mainframe computers. The result is that administrators are able to re-programme a communications infrastructure instead of re-build an infrastructure whenever they want to change or modify the services that run on top of it.

Contextually intelligent mobile technology and services are likely to accelerate adoption of programmable networks. As more powerful sensors and mobile devices capture environmental data that must be correlated across multiple data centres with stored user data, then analysed and returned to the user in real-time as a personalised, contextually relevant suggestion or action, the network infrastructure must become robust enough to handle communications traffic that exhibits both high volume and high complexity. [See sidebar on page 5]

Communications Services Providers (CSPs) are attacking the capacity and complexity problem along two broad fronts. To address high volume, they are deploying new radio interfaces and all-IP infrastructure based largely around the Long Term Evolution (LTE) standard. In addition, CSPs are pushing capacity deep indoors by nesting multiple large cell and small cell networks within one another—called heterogeneous networks (HetNets)—to load or offload user traffic onto the most optimal network infrastructure at any one time.

But handling high volume mobile data traffic is only one side of the coin. The complexity challenge for CSPs is equally acute as the architectural merging of the Internet and telecommunications


becomes increasingly mature. The transition to an all-IP future through LTE and similar technologies also enables new core architectures, such as Software Defined Networks (SDNs) and Network Function Virtualisation (NFV). These approaches abstract physical switches, routers and other components into a single virtual network layer that can be managed centrally. This allows CSPs to partition or ‘slice’ the same network infrastructure into application- or industry-specific functions. An SDN/NFV approach to architecture also enables communications networks to behave as platforms for service innovation at the CSP level and by third parties.

The end result is a telecommunications network that acts increasingly like a computing cloud.

Ubiquitous capacity

Whether voice or data, most mobile user sessions begin indoors or have some aspect of them that happen within a physical structure such as a building or an automobile. Moreover, people are bringing more sophisticated mobile devices and sophisticated expectations to their work, school and play. Not surprisingly, better and more powerful indoor mobile data capacity is an imperative for CSPs.

To meet consumer expectations of a seamless mobile experience, CSPs have started to assess both network performance and the quality of user experience as a function of app coverage more than voice coverage. App coverage measures the likelihood that a network will deliver sufficient performance to run a particular application at a quality level acceptable to the user. From the perspective of a CSP, app coverage is how well the mobile data network performs at the geographic edge of a cell. Partly in response to the need for better coverage for mobile apps, CSPs are installing new radio interfaces like 4th Generation (4G) LTE for higher capacity, and employing small-cell technologies to complement indoor and outdoor coverage.

In principle, LTE offers three broad benefits for dealing with mixed voice and data traffic. The first is a 10X increase in mobile data rates via Orthogonal Frequency Division Multiplexing (OFDM) compared to the 3G data transmission technology known as High Speed Packet Access (HSPA). Maximum mobile data rates according to the LTE standard include 300Mb/sec on the downlink and 75Mb/sec on the uplink. This increased speed also comes with increased spectral efficiency, up to 3X more capacity per bearer channel than a typical circuit-switched 3G network. Finally, LTE achieves one-fourth the latency (data packet transfer time from sender to receiver) of comparable 3G networks.

Along with adopting LTE in the macro-network, CSPs are expanding their use of HetNets for addressing high volume mobile data traffic in outdoor and indoor environments. HetNets attack the indoor/outdoor coverage problem along three broad fronts: the macro-cellular network provides wide-area broadband coverage from the outside; a dense mesh of enterprise, metro and small-cell technologies are linked inside for high-traffic areas such as airports or offices; and network intelligence steers traffic between macro and small-cell networks according to current demand to give the user a consistent indoor/outdoor experience [see graphic on page 4].

Two broad categories of small-cell technology underpin much of the current and future strategies for indoor radio coverage. Both are based on distributing antennas whilst concentrating base stations and backhaul. The current indoor coverage model is Distributed Antenna Systems (DAS). DAS architectures split the transmitted power of a single high-powered indoor antenna into a group of low-powered antennas over the same area. The placement of extra antenna elements, albeit at lower power, helps network designers work around

Whether voice or data, most mobile user sessions begin indoors or have some aspect of them that happen within a physical structure such as a building or an automobile. Not surprisingly, better and more powerful indoor capacity is an imperative for CSPs.

continues on page 6


0.1 Mbps

Figure 1: Anywhere a crowd gathers, so do devices and applications

Micro cell An auxiliary full-featured cell that can provide burst capacity for large indoor gatherings like a conference.

Macro cell Provides wide-area broadband coverage. Often based on 3G and LTE standards.

Indoor pico cell Takes over from the macro cell when the user moves inside.Wireless transmission

Links smaller cells to the rest of the network.

Wi-Fi Complements small indoor and outdoor hotspots with data and some voice capacity.

Heterogeneous networks provide data and voice capacity for complex, high volume traffic indoors and outside.

Source: PwC


Contextual services and capacity

Historically, network providers had a relatively simple idea of throughput. Throughput meant the average rate of successful message delivery through a given communications channel. More often than not, this meant the downlink side from the network to the mobile device. The volume of downloaded messages, the direction of message flow, and similar factors defined the networking industry, not simply in terms of the technology it adopted, but also the mentality with which it approached communications problems.

That model is changing rapidly as massive mobile data growth, a shift to visually oriented mobile data and the evolution to an app-centric usage model alter the mix of uplink and downlink mobile traffic patterns. Contextual services will combine and compound each force as end users generate large amounts of sensor data, captured media and applications that must upload rapidly to the network, process and then return almost immediately to the user as contextually relevant results.

At the most basic level, mobile communications networks are contending with unprecedented demand for capacity. Smartphones and connected devices pushed data past voice as the dominant source of mobile traffic in 2009. Ericsson reported in its June 2013 Mobility Report that total data traffic on mobile networks, in petabytes per month (1 petabyte = 1 million gigabytes), almost doubled between Q1 of 2012 and Q1 of 2013 (from just under 800 petabytes to just under 1,600 petabytes).1 Cisco’s Visual Networking Index also reported massive growth, calculating 885 petabytes per month of mobile data worldwide during 2012.2

These massive increases are happening as mobile video, image and visualisations are becoming the dominant mobile data types. According to Cisco, mobile video exceeded 50 percent of all mobile data traffic for the first time during 2012. As

1,3 http://www.ericsson.com/mobility-report

2 http://www.cisco.com/c/en/us/solutions/collateral/service-provider/visual-networking-index-vni/white_paper_c11-520862.html

same video destination during the same week compared to a user who did not experience a failure.4 Consequently, superior performance for mobile video is no longer a special case for high-end users but is expected as a standard feature by the general population of mobile data users.

The shift in mobile devices from person-to-person communications tools to the primary interactive lens through which users connect to people, data, applications and services is driving mobile data growth. Regardless of the particular instantiation of a contextually aware communications session, the need for low latency, high throughput and blended data and communications all point to fundamentally different capacity models and mobile network architectures than those which brought cellular communications into mainstream life.

4 http://www.akamai.com/dl/technical_publica-tions/video_stream_quality_study.pdf

screens for mobile devices grow in size and resolution, even more visual traffic is expected to travel over mobile networks. According to Ericsson, mobile visual data is expected to grow by 55 percent annually until the end of 2019.3 In 2013, the company reported that smartphone users who subscribed to music and video streaming services already consume more than 2GB of mobile data per month.

As important as the growth in mobile video volume, has been heightened user expectations for mobile video performance. For example, a 2012 study by Akamai Technologies and the University of Amherst on 23 million video streams from 6.7 million viewers showed that viewers start to abandon a video stream if it takes longer than two seconds to start, with each subsequent one second delay causing an additional 5.8 percent increase in abandonment. The study also showed that a viewer who experienced a failure in performance was 2.3 percent less likely to visit the

Sensors, captured media and applications upload data to the network.

The previous generation’s high-end performance is now expected by regular users.

Processing must often cross different network, database and service provider boundaries.

Users perceive network value as a function of application performance as much as voice quality.

Source: PwC


differences in material and architecture inside structures that can affect radio wave propagation. Optical fibre moves captured radio signals between a central facility and the remote DAS antennas. This makes sense in densely trafficked areas such as an airport or a convention centre.

Whilst DAS is currently a mainstream indoor coverage strategy, it requires considerable expertise and investment by the CSPs to deploy and operate. Other distributed antenna strategies aim to turn mobile broadband antennas and their placement into a near plug-and-play proposition. This is the design philosophy behind Ericsson’s Radio Dot System, which will launch commercially in late 2014. The actual active radio antennas or ‘dots’ weigh around 300 grams and deliver mobile broadband access to indoor users. Dots are connected and powered via standard Internet LAN cables that feed to floor-level radio units that all connect to a base station.

High capacity indoor coverage is not optional for CSPs that are faced with user expectations for instantaneous, reliable access to their apps and data wherever they are located. Equally important, the business case for LTE and small-cell radio coverage is being driven by the requirements of applications rather than voice. Along with providing more capacity, CSPs must provide smart capacity in order to prioritise different traffic streams and enable various business models. To make that happen, networks are becoming more programmable.

The network is a cloud

Communications networks are increasingly the front door to the customer for third-party service providers that comprise a larger portion of modern economies. At both the technical and business levels, CSPs will need to architect their networks to host ecosystems of third-party service providers engaging

end users with various contextual experiences, at different price points and under different business models. To meet the rapidly expanding virtual context of users and the businesses that support them, CSPs are moving to more programmable networks.

The shift toward making networks programmable starts with the all-IP architecture of LTE. In contrast to the circuit-switched model of cellular communications that led to today, LTE supports only packet-switched services. The goal of LTE is to provide seamless IP connectivity between a client device and the data packet network without disrupting the user’s applications whilst mobile.

Whilst LTE puts voice, data and applications onto a single delivery platform, new core network paradigms such as SDNs and its complement NFV are transforming how the network configures and operates on the inside. SDN and NFV use software to separate control of infrastructure elements from the underlying physical hardware to make a communications network operate more like a computing cloud. [See sidebar on page 7]

SDNs were pioneered in campus networking environments at the University of California at Berkeley and Stanford University in 2008. The purpose of SDNs is to allow administrators to shape communications traffic around different quality of service goals and/or business models without requiring admins to touch physical switches, routers or other hardware each time they want to make a change.

SDNs separate the part of network architecture that creates the map of nodes, links and addresses—the control plane—from the network architecture that makes decisions about what to do with inbound data packets (error correct, forward, reject, etc.)—also known as the data plane. This separation

Communications networks are increasingly the front door to the customer for third-party service providers that comprise a larger portion of modern economies.


abstracts the physical hardware from applications and services riding on top of the network. Network administrators can make changes or add and drop features from a central location instead of having to hand code hundreds or thousands of individual pieces of equipment. In addition to abstracting the network, SDN architectures support a set of APIs that make it possible to implement common network services such as routing, security, access control, bandwidth management, traffic engineering, quality of service and other forms of policy management, any one of which can be custom tailored to meet business objectives within their own organisations or on behalf of other organisations.

Whilst SDNs emerged in campus and data centre networking environments, NFV had its origins amongst European CSPs that did not want to continue buying proprietary network appliances to run each new telecom service. Instead, CSPs wanted to launch so-called virtual network functions to run on virtual machines housed on standard servers. The European Telecommunications Standards Institute (ETSI) launched the NFV group to spur the development of interoperable products to address diverse use cases.

In practice, NFV decouples various network functions, such as network address translation, firewalling, intrusion detection, domain name service, caching, etc., from proprietary hardware appliances, so they can run in software. NFV is designed to consolidate and deliver the networking components needed to support a fully virtualised infrastructure—including virtual servers, storage and even other networks.

The table below compares some of the key points of SDN and NFV.

The immediate effect of all-IP infrastructures combined with network abstraction architectures via SDN and NFV is a communications network that allows approved applications to instruct network elements directly about their needs (routing, security, and performance, for example). Conversely, the network can broadcast its capabilities, state, analytics and other data to applications that want or need to access them.

The ultimate impact of this bi-directional information flow enabled by SDN and NFV is a more open and platform-oriented approach to networking. Communications networks become configurable services

Category SDN NFV

Reason for being Separation of control and data, centralisation of control and programmability of network

Relocation of network functions from dedicated appliances to generic servers

Target location Campus, data centre/cloud Service provider network

Target devices Commodity servers and switches Commodity servers and switches

Initial applications Cloud orchestration and networking Routers, firewalls, gateways, CDN, WAN accelerators, SLA assurance

New protocols OpenFlow None yet

Formalisation Open Networking Forum ETSI NFV Working Group

Source: http://www.sdncentral.com/technology/nfv-and-sdn-whats-the-difference/2013/03/]

Table 1: Key features of Software Defined Networks (SDN) and Network Function Virtualisation (NFV)


Shopmart

Cars

Big corp

Figure 2: The network is the front door to the customer for many organisations

Each network slice is logically isolated with its own service level guarantees. This may span multiple data centres and network boundaries.

The network slice concept brings to telecommunications the same model of on-demand, elastic resource allocation associated with cloud computing.

that are accessed via APIs. Some embedded functions, such as firewalls, become customer controlled and tailored services in their own right. However, the same firewall service might also become part of a larger customer-facing bundle, such as a live video health counseling session, or a customer video conference with a tax professional.

The capability to provide à la carte network services based on application or business requirements rather than the specific configuration of physical infrastructure is called network slicing. In practice, a network slice is a logically isolated virtual network with its own service level guarantees that may span multiple data centres and network

boundaries. Communications traffic within a given network slice is logically isolated from other traffic and can be further enhanced with firewall and encryption technologies. Fundamentally, the network slice concept brings to telecommunications the same model of on-demand, elastic resource allocation associated with cloud computing.

Source: PwC


Application layer

Control layer

OpenFlow

InfrastructureLayer

Business application

Networkservices

Networkservices

Networkservices



API API API

Networks applications orchestrations & services





Controller platform

Base networkservicefunction






Service abstraction layer

Open DaylightAPI API API API

Data plane elements

Virtual switches Physical devices Virtual devices

A number of open-source initiatives are driving the development of Software Defined Networking (SDN) and Network Function Virtualisation (NFV). Two of the most prominent are the Open Networking Foundation (ONF) [www.opennetworking.org] and OpenDaylight [www.opendaylight.org].

OpenFlow is an effort by the ONF to define a standard communications interface that separates the control and data planes of a networking architecture. As a result, the network appears to higher level applications and policy engines as a single, logical switch. With SDNs based on OpenFlow, enterprises and carriers gain vendor-independent control over the entire network from a single logical point, which greatly simplifies the network design and operation. SDN also greatly simplifies the network devices themselves, since they no longer need to understand and process thousands of protocol standards but merely accept instructions from the Openflow’s SDN controllers.

SDN and NFV via OpenFlow and OpenDaylight

Like OpenFlow, the OpenDaylight approach to SDN and NFV networking is a three-tier stack. The top level is called the Network Apps and Orchestration layer, which consists of business and network logic applications that control and monitor network behaviour. The middle layer, called the Controller Platform, is the framework in which the SDN abstractions can provide a set of common APIs to the application layer (commonly referred

to as the northbound interface) whilst implementing one or more protocols for command and control of the physical hardware within the network (the southbound interface). At the bottom layer are the physical and virtual devices, switches, routers, etc., that make up the connective fabric between all endpoints within the network.

Source: https://www.opennetworking.org/sdn-resources/sdn-definition

Source: www.opendaylight.org

Another important SDN standards player is OpenDaylight, which emerged from the Linux Foundation. OpenDaylight is an open-source project with a modular, pluggable, and flexible controller platform at its core. This controller is implemented strictly in software and is contained within its own Java Virtual Machine. As such, the controller can be deployed on any hardware and operating system platform that supports Java.

Officially, the two initiatives are not in direct competition. OpenDaylight will include support for the OpenFlow protocol, but will also be extensible to support other emerging SDN open standards, according to the OpenDaylight Foundation. The fact that many of the leading infrastructure and service provider organisations are founding members of both groups supports the case for complementary development.


At the business level, this means that CSPs can configure slices of the same network customised for different applications or industries. Application developers are able to access network services and capabilities without being constrained by the details of implementing them in physical infrastructure. In that sense, the game changer from being able to slice a network is the evolution of communications infrastructures from being just utilities to becoming true innovation platforms. According to Ulf Ewaldsson, Chief Technology Officer for Ericsson, network operators and third-party service providers will be able to access and direct networking resources under a cloud-based model similar to how they access storage and computation. “It means that developers can quickly build and deploy services outside of the operator’s domain that are using the capabilities of networks from the device all the way to the data centre.”

It’s an app-driven world

The rise of virtual context means that CSPs must now assume that every communications device is simultaneously a computing and sensing endpoint for a user’s personal cloud. That personal cloud, in turn, interacts with the surrounding physical environment, creates a user’s proximity network and enables a portfolio of CSP and third-party clouds that deliver contextually relevant and intelligent experiences.

These contextually intelligent services will have significant uplink traffic as users generate massive amounts of data through device and environmental sensors, video and audio capabilities that were covered in the previous article. Aside

from the low-latency required to provide the user with a quick and seamless experience, the network must route data traffic through the personal cloud of the user plus multiple clouds of various service providers depending on the user’s current situation.

Consequently, mass-market expectations about network coverage and capacity for applications are matching the expectations that were previously restricted to high-end users. This suggests that the value proposition and competitive centre of gravity of communications networks will increasingly revolve around the needs of applications rather than voice or messaging, not to mention the value propositions and business models of third-party service providers.

Thus, virtual context is a user-driven rather than an infrastructure-driven technology landscape. The performance characteristics and value-creating activities of networks will evolve increasingly according to the requirements of users. In that sense, the expanding virtual context of users and the evolution of the networked cloud go hand-in-hand. In the next article in this series, we will explore the final layer to the contextual stack, the intelligence engines in the cloud that take data from a user’s physical and virtual context and analyse it for predictive actions or suggestions.






Pierre-Alain Sur Global Communications Industry Leader PricewaterhouseCoopers LLP [email protected]




Mobile Innovations Forecast Sensing and making sense: Device and environment underpin contextually aware services

The core building blocks for contextual awareness and intelligence emerge from understanding the immediate situation of an end-user. Knowing a user’s physical location, time-of-day, calendar and data associated with movement of the mobile device captures a large swath of information required by predictive algorithms that make contextually relevant suggestions or automatically execute actions on behalf of the user. This situational device data, in turn, will be augmented by visual and aural information to develop more nuanced, semantically rich descriptions of a user’s current environment and likely intent.

To accomplish this, the technology industry is already in the midst of a race to connect mobile devices and their physical environments with sensors, beacons and other data gathering and broadcasting technologies. Not just smartphones and tablets, but fixed locations and everyday objects are gaining the ability to communicate wirelessly with each other and with the end user. PwC expects this race to accelerate.

This article focuses mainly on physical context information harvested and packaged by accelerometers, gyroscopes and other mobile device sensors that create a picture of the state of the device. Cameras and microphones are


By Raman Chitkara, Global Technology Industry Leader

Phase II synopsis

This is the second article in the Mobile Innovations Forecast Phase II: New Technological Capabilities. Here’s a roadmap of the series:


- This second article focuses on how device and environmental sensors interact to capture information to model the user’s physical context.

- The upcoming third article will examine the communications network that will enable information from the physical environment to be correlated in the cloud to create the virtual context layer of the user experience.

- The fourth article will describe the modeling and analytics, mainly in the cloud, that will enable the mobile device to become as intimate as a close friend, if you allow it.





also significant sensors. However, their utility for generating a contextual model of the end user is tied more to image and audio recognition applications that live in the cloud. Analytic capabilities that make sense of captured images or audio will be explored in the fourth article of this series. (see Synopsis)

Most of the low-level technology ‘conversations’ between the physical world and the user’s mobile devices will take place at the sensor level. In addition to continuing improvements

in sensors themselves, new capabilities for understanding physical context are evolving in these areas:

• dedicated processors for contextual awareness

• sensor fusion, which turns data from multiple sensors into usable information

• new frameworks for security in peer-to-peer or ad hoc networking

• geo-fences, which are virtual boundaries for physical areas.

Warehouse 1—Item Locations

Order Number: 1234599-NP2345

Items List Shipping Location

Zone ADamaged

Aisle K1

Aisle K2

Tabl

e 1-

KTa

ble

2-K

Overflow

Aisle K3

Aisle K4

Aisle K5! ALERT—Additional item pick-up requestOrder: 2345678-ZA9876

X1 2

43

4 5

32

1

Indoor navigation optimised through sensors

Figure 1: The interactive warehouse


Mobile device sensors already respond in rudimentary ways to changes in a user’s physical context. Turn a smartphone or tablet from portrait to landscape orientation and the display automatically refits an image. Bring a smartphone to your ear and a proximity sensor tells the main processor to shut off the touch screen. But these limited, primitive examples of individual sensors adapting a mobile device’s behaviour around a user’s context are giving way to multiple sensors working together to paint a rich picture of a user and his or her environment.

This evolution is in line with the Introduction to the Mobile Innovations Forecast Phase II: New Technological Capabilities, which asserts that mobile devices are evolving into contextually smart digital assistants. A big part of that transformation involves enabling mobile devices to understand the physical situation of the end user and employ that knowledge to serve her needs, often without requiring the person to state them explicitly.

Practically speaking, the contextual messages being communicated about physical context by people, objects and locations can be boiled down to “This is who and/or what and where I am right now, and based on my calendar information where I expect to be in the near future.” ID signals and beacons within user devices or embedded in the outside world or places and objects capture and communicate who and/or what. Device and environmental sensors capture and communicate location and activity as it relates to use of the device or mobility of the user.[See Figure 1]

As more sensors spread into more devices, places and objects, the smartphone is emerging as the core interface between a sensor-connected world and those who live in it. The physical state of mobile devices and the user’s immediate environment are the bedrock of mobile contextual information and the most common starting points for building contextually aware services.

1 For more information go to https://developer.apple.com/library/ios/documentation/CoreMotion/Reference/CoreMotion_Reference/_index.html

Device sensorsToday’s smartphones already contain multiple sensors that generate various types of device and environmental data. Standard components typically include two cameras front and back; two or more microphones; an accelerometer to measure acceleration; a gyroscope to measure orientation; a magnetometer (compass); an ambient light sensor and a proximity sensor.

The purpose of sensors is to create an accurate, robust depiction of the position of a mobile user in physical space through his device, time-of-day and how that relates to his current calendar or situation; and his proximity to other devices, services, objects or locations. Applications can access and mine this data to adjust automatically or return information or suggestions based on where the user is and what he is doing.

This is fueling an innovative push by mobile technology OEMs to launch a new class of processors dedicated to capturing and packaging sensor data. The Motorola X8 Mobile Computing System and the Apple M7 are two early examples.

The Motorola X8 Mobile Computing System includes a contextual processor and a natural language processor in addition to its main CPU. To save battery power, the X8’s contextual processor might work with a device accelerometer, gyroscope and ambient light sensor to detect whether a mobile device is in the user’s pocket, in a bag or lying face down. The main processor uses that data to light up the display only when the user needs it. The contextual processor also feeds into more dynamic situations, like when a user is trying to take an action shot with the camera. It uses accelerometer data to detect motion, checks ambient light and proximity sensors to determine if the phone is out of the user’s pocket that suggests the user is ready to shoot.

The Apple M7 is a separate processor (or coprocessor) announced as part of the iPhone 5S launch. Like the Motorola contextual processor, the M7 builds a sophisticated motion model of the user without requiring resources from the device’s main processor. Along with saving battery life, the M7 enables developers to pull motion sensor data into their applications through its CoreMotion API1.

Power management is just one example of the new generation of motion sensor-focused processors. Another important function is indoor navigation. Whilst communications infrastructure and services providers continue to build out network coverage indoors (more about this in the next article), motion technology providers are using device sensor information to derive more accurate indoor directions for users.

Regardless of its ultimate purpose, sensor data must be normalised and packaged if it is to be used by higher-level applications like fitness or indoor navigation. A big part of this involves filtering out environmental noise. For example, the compass (magnetometer) in a mobile device contends with magnetic anomalies when a person enters an elevator or rides an escalator. These activities slightly change how the mobile device interprets magnetic north.

There is a growing cadre of start-ups and established chip vendors fielding hardware and software solutions for improving the accuracy of sensor data by filtering environmental noise and combining multiple sensor inputs for contextually richer and more meaningful data for use by applications. [See Figure 2]



The capability to turn raw sensor data into something usable by applications is called sensor fusion. Sensor fusion is device-resident software that combines sensor data or data derived from disparate sensor sources to produce information that is better in some way than is possible with individual sensor sources alone. Better might mean more accurate, more complete and more dependable, or an

emergent view generated by fusing the results of several distinct sources of sensor input. Optics offers an early example; specifically, stereoscopic vision, in which a computer calculates depth information by combining 2D images from two cameras set at slightly different viewpoints.

Dan Brown, CEO of Sensor Platforms, a venture-backed company in Silicon Valley, says the goal of sensor fusion is to create a ‘confidence engine’ that aggregates, normalises and packages sensor data into a form usable at the application layer. “Let’s say we detect movement of the device,” he says. “Is there movement because the user stood up, or because the user moved the phone from one side to the other of his jacket? Or did he pick up the device from a table? Sensor context awareness identifies those movements and the confidence engine sends that information up to the application. The application is only responding to the information that we are giving it, and we’re not bombarding and waking up the application processor each time there’s an update.”

Enabling better system performance whilst making situational data available and useful for applications incentivises OEMs to pack more sensors into almost every networked device with the smartphone acting as the primary sensor hub. The other side of the coin involves instrumenting the outside environment with sensors, tags, beacons and other targets that will interact with mobile devices to make a sensor-derived personal cloud of information and services closer to reality.

N

Applications

Sensors

Contextualprocessor

Contextual

OS

Low-level data is transformed for human benefit

Figure 2: From input to action


The outside world becomes the desktop

Sensors are not restricted to mobile devices. The diffusion of sensors and proximity technologies into physical locations and objects is equally rapid. Service providers are instrumenting public environments, ranging from sports stadiums to shopping malls, with sensors, beacons, tags and other radio-connected computing nodes. The goal is to enable the user to share information with her proximate environment and get something in return.

An example is a coffee shop that recognises a regular customer when she’s within 10 metres of the front door based on the ID signature of her smartphone or fitness tracker. Armed with that knowledge, the coffee shop might pre-order the customer’s favourite beverage.

Beacons and ID

A significant problem with indoor navigation is the line-of-sight requirement for most GPS technologies, which often proves difficult inside shopping malls and other buildings. At the same time, new indoor navigation efforts using low-power radio beacons within buildings are gaining scale. Beacons are small, wireless sensors that are placed within a physical space and transmit an ID code to announce their presence to a compatible mobile device and, in so doing, establish a digital perimeter.

A user’s mobile app can be enabled to look for a beacon’s ID transmission and use that information for better navigation to a target area in an indoor space and for triggering a notification of location-relevant content, offers and promotions. Qualcomm introduced Gimbal proximity beacons in 2013 to complement GPS by allowing devices and applications to derive their proximity to beacons that continually broadcast an ID code twice

a second. In September 2013, Apple announced its iBeacon indoor positioning system. Like Qualcomm, iBeacons are low-power, low-cost transmitters located within an indoor structure. When an iOS7 or Android user enters the transmission perimeter established by an iBeacon, push notifications can be sent to the device whilst the physical location can track entry and exit data generated by users interacting with the beacon.

The Qualcomm and Apple beacons employ Bluetooth Smart, also known as Bluetooth Low Energy, to enable their systems. In contrast to Near Field Communication (NFC), which requires that a compatible device and target be within 20 centimetres of each other to trigger an event, Bluetooth Low Energy can reach up to 50 metres in distance. It is still too early to determine whether NFC or Bluetooth Low Energy will prevail in the battle to navigate the great indoors.

A key capability for making the outside world responsive in this way involves creating peer-to-peer (P2P) connections between sensor-equipped mobile devices and physical venues or objects. A significant effort involves the use of physical proximity ‘beacons’ that are embedded in physical locations to connect P2P to mobile devices to trigger a processing event. [See sidebar, Beacons and ID, below]

Along with physical beacons, the core networking technologies, such as Wi-Fi and Bluetooth, are already widespread, with newer technologies such as LTE Direct on the way. However, data transmission is only one part of a larger palette of capabilities required to remove friction in P2P networking. Developers trying to build applications must contend with device and service discoverability, security, different radios or platforms, pairing protocols and the like to make ad hoc device and service connections easily accessible to the end user.


Effect: House locks doors, arms alarm systems and resetstemperature to away levels

Geo-fence: Home security

Effect: Informs security of arrival,boots up computer, removes forwarding of phone

Geo-fence: Work company

Effect: Informs security of departure,locks computer, forwards phone, activates reminder to pick-up child

Geo-fence: Work company

Effect: Confirms parent id, oks child pick-up

Geo-fence: School

Effect: Sends message throughsmartphone inquiring if he would like usual order

Effect: Alerts sent to smartphone,reminder of oil change need

Geo-fence: Car dealership

Geo-fence: Coffee house

Crossed boundary: 7:42 am

Crossed boundary: 4:35 pm

Crossed boundary: 4:22 pmCrossed boundary: 8:12 am



Figure 3: Macro-map and micro-tasks

Maps become geographic references plus planning tools


Amongst evolving new capabilities are some of the first frameworks for handling the higher-level challenges of P2P networking. AllJoyn is an open-source project launched by Qualcomm that provides a universal software framework and core set of system services to enable interoperability amongst connected products and software applications to create dynamic proximal networks. Its core building blocks and services address discovery, connectivity, security and management of ad hoc proximal networks amongst devices that can interact regardless of how they are connected—Wi-Fi, Ethernet, Powerline, etc.

Personal cloudProximity technologies like AllJoyn, which fit on top of established networking protocols like Wi-Fi Direct (used by Apple AirDrop) or Bluetooth, aim to create a new kind of cloud—a personal cloud of devices and applications around the user wherever he goes. It’s a dynamic environment that evolves as the user moves through the physical world. Most likely managed by a smartphone, a constellation of mobile devices owned by the user will tap into that proximal cloud to access information and services.

Dr. Paul Jacobs, the CEO of Qualcomm, refers to this personal cloud and the capabilities it unleashes as a ‘digital sixth sense’ because it uses all the sensor data as interpreted and extended by analytics to augment the human’s five senses.

Whilst proximate capabilities like AllJoyn make ad hoc networking possible for devices, other new capabilities like geo-fencing aim to draw dynamic, virtual perimeters around physical locations of any size, from office buildings to entire neighborhoods. A geo-fence is a digitally created border around a geographic point. When a location-aware device enters or exits a geo-fence, the

device receives a notification. Originally designed for child location services, geo-fencing leverages awareness of a user’s current location with awareness of nearby features that may be of interest based on a user’s stated preferences or inferred from her past history.

To create a geo-fence, a service provider specifies latitude/longitude of the target location and then specifies a radius from that point to adjust a proximity filter that will determine whether a user receives a notification. A geo-fence can be dynamically generated—as in a radius around a store or point location. Or a geo-fence can be a predefined set of boundaries, like school attendance zones. Service providers can have multiple active geo-fences and even nest them within each other. They can also specify an expiration time and date for the geo-fence.

On either side of a geo-fence, a user or a venue owner can employ identity beacons that trigger an action when a border is crossed. Some of these beacons are actual physical devices, such as Qualcomm’s small blue keychain beacon, a simple device that uses Bluetooth Low Energy (BLE) to send out an ID signal twice a second. A contextual application running in the background on a mobile phone or another device will send the received token ID to a server, which can send information back to the phone. The specific information payload depends on the application and the behaviour to which it responds.

Some of the first trials of identity beacons and contextual services have already been run in sports venues such as AT&T Stadium, home of the NFL’s Dallas Cowboys. Other public spaces, such as conference arenas, are experimenting with ID beacons to determine where and for how long people visit specific geo-targeted areas such as concession stands or parking garages.

Geo-fencing is expected to expand beyond just navigation and alerts to include more direct transaction activity. This is because most ‘real’ commerce still does not take place online, according to Jeff Miles, VP of Mobile Transactions at NXP, a semiconductor company. “The number is still less than 10 percent of the total transactions are what we would call pure online transactions, think of payment transactions,” he says. “That means a full 90 percent still happen at a physical point of sale. So the ability now to interact with those touch points with an electronic device, that’s a huge opportunity of how you can affect those transactions.”

The spread of geo-fences, beacons, Near Field Communication tags and other sensor-based devices and services is really about connecting the power of


the mobile device and the cloud with physical places and objects. Some call it the Internet of Things or even more grandly, the Internet of Everything. Regardless of word choice, in short order nearly each and every connected device or location will become a networked computing node, a development with staggering implications for the mobile technology and services ecosystems.

Conclusion: Signal or noise?We want two things from any contextual model. The first thing we want is a filter. We want technology and services to remove everything, except the information that we really need to accomplish our objective. We want to find our way outside or indoors. We want to know how many steps we took or calories we burned. But we don’t want a dead battery by the middle of the day. We don’t want to miss an important call because of a game or a retail offer.

Sensors and intelligent processing of our immediate physical situation help filter out as much background noise as possible. Understanding our physical context is also important for maintaining the performance of devices. But that’s only half of the equation.

After we filter out the noise, we want to make sense of the signal. We want to apply intelligence to the filtered information that is important to us. Getting to that point requires that we build on top of our physical context to include stored information, preferences, applications and services in the network and in the cloud.

Granted that today’s mobile technologists have a good idea for making that happen, there remain larger unsolved problems around capturing physical context and making it valuable for the end user. Some of these problems are technical, such as incorporating machine vision into the portfolio of sensors capturing contextual data or determining the optimal standards for peer-to-peer communication between sensor-equipped mobile devices and objects and locations.

However, the more important innovations will be social and business related. For example, how should service providers alert and engage users with a geo-fence smoothly and quickly? Of course, there are technical details for making that connection, but the process starts by convincing the end user of the value. Moreover, the notion of privacy in a contextual age is no longer a binary decision, but comes in degrees and layers. Creating the interfaces that allow end users to adjust their privacy settings as easily as they change the volume on their devices or search key words is still a work-in-progress.

The next few years will bring the merging of physical space and digital information/services into sharper focus. As we make our mobile devices and our physical environments more intelligent, our basic definition of interactivity must expand to include how people access and manipulate digitally enabled locations and objects, not just web pages. The irony, of course, is that linking the physical world to the digital world to expand interactivity requires a transformation of telecommunications as profound as that unleashed by the original Internet. Networks and the end user’s virtual context will be the focus of the next article in this series. Stay tuned.






Kayvan Shahabi US Technology Advisory Leader PricewaterhouseCoopers LLP [email protected]




Mobile Innovations Forecast Phase II Introduction The magic of advanced technology

Growth in contextual awareness capabilities will transform mobile devices into digital assistants. Science fiction writer Arthur C. Clarke’s Third Law states: “any sufficiently advanced technology is indistinguishable from magic.” While people often focus on the word ‘magic’, the key phrase is actually ‘sufficiently advanced technology’. As we move into Phase II of PwC’s Mobile Innovations Forecast (MIF), we are now entering the realm of sufficiently advanced technology.

Until Apple’s iPhone debuted in 2007, mainstream mobile innovation emphasised how well a mobile device communicated. Sufficiently advanced technology was defined according to the price/performance of placing phone calls, sending messages or displaying simple text or graphics. Four basic applications—phone, messaging, contacts, camera—provided the bulk of the end-user’s mobile experience. The structure of the mobile industry revolved around the coverage, quality and price of cellular networks. Mobile devices lagged behind desktop PCs in storage, processing power and data handling capabilities.

However, the iPhone put mobile innovation on a completely different track. Since 2007, sufficiently advanced technology has referred to how well a mobile device computes. According to Daniel Eckert, Managing Director, Emerging Technologies at PwC, the innovative focus of mobile computing involves integrating a triad of communications, applications and

sensing platforms. “These three platforms ‘converse’ with the user in a continual loop of personalised applications that draw information from a user’s physical, virtual and social environments,” he says.

Daniel Eckert

Managing Director, Emerging TechnologiesPwC

As this process iterates, mobile devices and the supporting wireless infrastructure become far more than advanced communications tools. They become an extension through which a growing number of people participate in modern culture.

Compute-centric mobile participation goes beyond the idea of more people using more devices to communicate more frequently. Participation starts at the API level as mobile devices exchange information drawn from in-device sensors, plus information stored in the cloud or embedded in physical objects or landmarks. Participation also occurs at the user profile level, with mobile devices allowing people to broadcast preferences, intent or even telemetry about their physical bodies, location or social graphs. Finally, participation takes place at the aggregate level as more users with more powerful devices create and scale the feedback loops that power technical, economic and social ecosystems.





Physical

SocialFacebookLinkedInTwitterTumblrPinterestFoursquare

Cloud storageApplicationsID & walletCalendarsContactsNetworking

Virtual

Smart glasses� captures image, video and audio� scans coded markers� enables multimedia chat� updates social networks� syncs with other mobile devices Wellness monitor

� processes exercise data (steps, reps etc.)

� analyses nutrition/ calories of grocery and restaurant foods

� analyses perspiration for chemical markers

� syncs with wearable exercise shoes/clothes

� syncs with personal health portal

Health monitor� captures resting/active

pulse, BP, temperature� analyses cholesterol,

insulin and similar markers

� syncs with wellness monitor for building health profile

� communicates with healthcare provider

� offers suggestions for health improvement

Mobile device� primary mobile information hub� wallet and credential hub� application management hub� media and communications input hub � interface between user and core

service portfolio

Networked wearable devices� records physical activity for

upload to wellness monitor� records physical activity

to a time log� analyses physical performance

against goals� displays progress/regress

against goals� syncs with opt-in social networks GPS

CompassWeather

AccessoriesHardware Peripherals

Sensors OSTraffic

The current situation

The first phase of the Mobile Innovations Forecast (MIF) introduced a quantitative model that analysed the rate of improvement in key technologies considered fundamental to mobile innovation. Readers of MIF Phase I have followed a steady stream of data and analysis regarding the likely trajectories of device and infrastructure connectivity; application processor speed; DRAM and storage improvements; as well as image sensor, display and mobile operating systems.

The rapid improvements in price and performance of these various technologies suggest that mobile innovation has become self-accelerating; that is, the results of each advance enable further advances to develop even more rapidly. But to what end are these innovations driving?

The next phases of the Mobile Innovations Forecast (New capabilities, New use cases, New business models) will attempt to answer that basic question.

Figure 1: The contextual man

Source: PwC


Given such profound technical and behavioural shifts, what knowledge of new capabilities is needed by the technology, telecom and media sectors to engage users for whom mobility is not just a physical fact, but also a state-of-mind? This is the core issue addressed by the next phase of the PwC Mobile Innovations Forecast (MIF). “Phase II, New capabilities,” will identify and analyse the new technological advances—made possible by the enabling technologies covered in Phase I—that generate contextual intelligence from mobile users’ physical, virtual and social environments.

PwC forecasts that the next phase of mobile innovation will revolve around capturing and modeling the contextual situation of mobile users. Such knowledge will become the primary resource for predictive mobile applications and services that will address mobile users’ needs and desires in near real-time, and often before the users themselves reveal what they want.

As we enter Phase II of the MIF, PwC forecasts that the following new capabilities—each of which will be the topic of a future article—will form the basic architecture underpinning contextual awareness and intelligence in next generation mobile devices, networks, applications and services:

• Location and navigation

• Device sensors and user interfaces

• ID and security technologies

• Next-generation networks and clouds

• Mobile operating systems

These categories have always been part of the basic mobile technology stack. However, given the shift from communications to computation as the

driving paradigm for mobile innovation, PwC expects that the purpose and nature of these technical categories will change dramatically.

This forecast exists within PwC’s framework for understanding various dynamics affecting the broader technology sector. Mobile innovation is one of four market forces that are redefining customer demand, expectations and business opportunity for technology companies. The others are cloud computing, social technology or media and the emergence of intelligent devices. Individually, each is turning the rules of the broader technology sector upside down. Collectively, they are co-mingling in ways that paint a forward-looking picture that is starkly, even radically, unlike the past.

Context and companionshipThe goal of Phase II of the MIF is to analyse which new technical capabilities help make mobile systems more context-aware, and which new capabilities help add contextual intelligence to users’ interaction with their physical, virtual and social environments. Contextual awareness means that mobile devices are able to capture and analyse users’ relationships to people, organisations, places and objects around them in the broadest sense, including but not limited to the proximate physical environment. Armed with such knowledge, a contextually intelligent mobile device or service can infer a user’s needs, desires and even intentions without requiring the person to state them explicitly.

Contextually intelligent mobile services are fueled by data drawn at multiple levels. Some data streams are device, network or application-centric, while others are bound closer to individual

Our research hypothesis is that new mobile capabilities to 2016 will enable mobile devices and services to become contextually aware and intelligent about end users in order to help them participate in modern life.


users, their relationships and their activities. The current consensus is that location and navigation data rank as first amongst equals within the palette of mobile contextual inputs for both opportunity and challenges. Other critical contextual inputs include user ID, device sensor data, data generated by networks and computing clouds as well as individual and social activity data by the user.

When thinking about contextual capabilities, it is useful to imagine the mobile device behaving more as a digital assistant than a communications and Internet access tool. A digital assistant tries to understand a mobile user’s contextual situation to infer intent and to offer suggestions or services that help accomplish the user’s goals.

One of the first steps for a digital assistant is to understand a user’s immediate physical environment. Where is she located? Is the user stationary or moving? Is the device in a user’s pocket, in a purse or in her hand? What other relevant landmarks or objects are nearby based on a user’s prior experience or preferences? Who is nearby in the user’s social network?

Such questions and many others help a digital assistant build a detailed simulation of the user’s world. But contextual awareness and intelligence doesn’t end with the digital assistant mapping a perimeter of interest points and objects around a user. A digital assistant must also utilise predictive models of the user’s relationships to her applications, personal or preferred

My Tracker

Daily calorie limit: 2,000Calories remaining: 650

Warning!

Estimated calories: 1,064

Your meal:

365

459

240

Over daily limit

Over daily limit

What’s the caloriecount on a burgerand fries again?

How many calories do I have left??

800?

300?

Figure 2: A digital assistant

Source: PwC


information as well as her social world. What choices did the user make in the past in a similar situation? What are the user’s stated preferences for this condition? What can be inferred? What did the user’s friends or colleagues do in a similar situation?

The ultimate purpose for building models of observable entities such as environment and behaviour, and inferred entities such as relationships and intentions, is to help users optimise the choices they make. For any technology to succeed in adoption and use, it needs to empower a user to drive to a desired state. Consider health and wellness.1 “Mobile digital assistants can provide patients with the information and immediate feedback loop they need to direct and reinforce desired behaviours,” according to Chris Wasden, Managing Director and Global Healthcare Innovation Leader at PwC.

Chris Wasden

Managing Director and Global Healthcare Innovation LeaderPwC

For example, Japan’s NTT DoCoMo launched a ‘wellness phone’ that includes a pedometer and health monitoring software. Data drawn from sensors in the phone is analysed and insights provided to the user along with advice

on healthy lifestyle choices. Such a dialogue enables mobile devices and services to move beyond just utility functions (e.g., taking pictures of the food you eat) and move increasingly toward having an aspirational or lifestyle-centric relationship with users such as signaling when they are near their daily maximum amount of consumed calories.

PwC believes the ability for a digital assistant to join contextual information with a direct feedback loop distinguishes next-generation mobile devices from today’s smartphones. We predict that more than simply storing data or processing applications, contextually aware mobile devices, applications and services will continually ‘learn’ about their users to provide better experiences the more they are used.

Context as a technical conceptAccording to researchers at the Georgia Institute of Technology,2 context is any information that can be used to characterise the situation of an ‘entity’— a person, place or object—that can be considered relevant to an interaction between a user and an application. For example, if a location sensor on a smartphone or tablet detects that a user is either in the United States or Canada, the search results in terms of distance to a point of interest can be expressed either as miles or kilometres. The

1 The rapidly evolving mHealth sector will be the focus of a future article in Phase III New use cases of the Mobile Innovation Forecast

2 Dey, Anind, and Gregory Abowd. “Towards a Better Understanding of Context and Context-Awareness” HUC ‘99 Proceedings of the 1st international symposium on Handheld and Ubiq-uitous Computing. (1999): 304-307. Web. 28 Nov. 2012. <http://smartech.gatech.edu/xmlui/bitstream/handle/1853/3389/99-22.pdf?...1>.


More becomes different

The most recent data from Cisco’s Visual Networking Index Mobile Forecast suggests that we are well advanced into a smartphone-centric paradigm of mobile data usage.

Smartphones comprised less than 20% of total global handsets in 2012, but they were responsible for 92% of mobile data traffic around the world. According to Cisco, the typical smartphone in 2012 generated an average of 342 MB of data traffic per month, some 50X higher than a standard feature phone.

Cisco projects that mobile data traffic volumes will grow at a compound annual growth rate (CAGR) of 66% over the next five years to reach 11.2 exabytes

per month worldwide by 2017, a 13X increase from 2012. There will be over 10 billion mobile-enabled devices by 2017, a combined figure larger than the world’s projected population of 7.6 billion human beings.

Such a dramatic scale is changing the nature and scope of mobile innovation. New mobile capabilities on the supply side and broad, diverse markets on the demand side have started catalysing each other to grow faster and larger with each iteration. But even if we accept that mobile computing has become increasingly pervasive and self-accelerating, there remains the question—accelerating to what end?

66%

11.2 exabytes

Mobile data traffic will grow at a compound annual growth rate of

to reach

20172012

16%

34%

other handsets

smartphones

per month by 2017.

Source: Cisco and PwC estimates

http://www.cisco.com/web/solutions/sp/vni/vni_mobile_forecast_highlight/index.html

http://www.cisco.com/web/solutions/sp/vni/vni_mobile_forecast_highlight/index.html


location data point not only describes itself (longitude and latitude), but it also becomes a contextual input to the display function of a mobile application (miles or kilometres).

Drilling further into how technical systems handle contextual inputs, another definition by Forrester Research3 states that context-aware technology programmatically determines the use condition of itself and its user, and then adapts its own features and behaviour based on historical and current conditions, behavior, preferences and circumstances.

Although both approaches provide a formal description of context for computing environments, they are limited by their attempt to capture as much as

possible in a single definition. Rather than focusing on a single declaration of what context is, PwC believes that contextual intelligence in mobile computing emerges via predictive models that draw situational information from three main environments:

• The device environment (e.g., available power, OS, network, processing, storage, etc.)

• The physical environment (e.g., location, weather, lighting/noise levels, codes attached to physical landmarks or objects, etc.)

• The user environment (e.g., ID, applications, stored data, preferences, activity history, social connections, etc.)

These three categories characterise a

Contextualintelligence

User

DeviceEnvironment

Predictive models draw information from three main environments.

3 Ask, Julie A. “The Future of Mobile eBusiness is Context”, Forrester Research. 1 May 2012.

Figure 3: Sources of contextual data

Source: PwC


Location is Ground Zero

It’s clear that one of the most important contextual inputs is a user’s physical location. Location has become a major driver for mobile application development. The providers of the main mobile operating systems (OS) such as Android, iOS, Windows and Blackberry, are fielding sophisticated APIs for app developers to pull location-based data into mobile apps. Giving developers better access to more granular location information will increase the richness of human-computer interaction and will open up opportunities to add value to more sophisticated user demands, especially navigating indoors.

Indoor navigation is both a valuable application in its own right plus a contextual input into other mobile apps and services. Until recently, the price/performance of device sensors was a limiting factor, along with less advanced mobile OS and development environments. However, smartphones and tablets now contain more sensors than before.4 The fusion of data from sensors such as GPS, compass, accelerometres, barometres, temperature and pressure gauges, promises better indoor wayfinding as well as scope for painting digital information onto physical landmarks and objects to enable new services such as mobile augmented reality (AR).

Popularised by new form factors such as Google Glass, mobile AR merges real-time digital information into a user’s literal field-of-vision. The user simultaneously experiences physical reality with a digital overlay of information. The device displays information based on either image recognition of a specific target such as a Quick Response code (marker-based AR) or the more common version of mapping a device’s location to a database of information targets attached to physical coordinates (markerless AR).

Sensors to orient a mobile device in 3D physical space underpin mobile AR and a host of other application categories such as games. During these experiences, mobile users aren’t ‘pointing’ their devices at target locations or objects so much as they are ‘flying’ their devices. The three main spatial coordinates (width, length, altitude) are joined with three orientation angles (pitch, yaw and roll). Spatial coordinates are supplied by GPS, compass and barometre, whilst orientation data is furnished via accelerometres and gyroscopes.

The possible number of use cases and business models unleashed by better location and sensor data is huge and growing. But so is concern over user privacy in the face of such detailed tracking. The upshot for decision makers is that contextually intelligent mobile services require near equal focus on both technical and social engineering.

Location has become a major driver for mobile application development.

4 Device sensors and their applications will be the topic of a separate article in this Phase II series. Source: PwC


mobile participant’s situation regardless of whether that participant is another person, a place, an object, an application or a service. Depending on the expressed or inferred goal of the user, contextual information from one or more of these environments will be accentuated.

Going forwardThis article launches Phase II, New capabilities of the Mobile Innovations Forecast. We will move up the innovation stack not only in coverage, but also methodology. Whereas, the Phase I articles on mobile innovation analysed quantitative data drawn from the Mobile Technology Index, this series of Phase II articles will take a more qualitative approach based on deep interviews with mobile innovation thought leaders, both within PwC and the industry-at-large.

In terms of what readers can expect from these articles concerning new capabilities, we will build our analysis of context as a driving force from the ground up. There will be a series of six follow-on articles, the first of which will drill deeper into contextualisation as a technical and organisational concept. The

second article will explore location and navigation as contextual inputs. A third article will examine the role of device sensors and new user interfaces that capture contextual information about the user and her surroundings. A fourth article will focus on ID and security issues affecting people, places and objects in this rapidly evolving environment. The fifth article will highlight how next-generation networks and computing clouds allow contextual data to be stored and accessed in real-time. The sixth and final article will look at how contextually focused services will impact mobile operating systems.

Granted there are numerous sub-domains to explore, but we feel confident that context-rich mobile experiences represent the future of mobile innovation. This isn’t simply because contextual awareness and intelligence are now made easier through the sufficiently advanced technology identified in Phase I. Instead, we believe that the lasting impact of contextual technology lies in its ability to enable mobile technology to act more human. When that process works well, it does feel like magic.


© 2013 PwC. All rights reserved. PwC refers to the PwC network and/or one or more of its member firms, each of which is a separate legal entity. Please see http://www.pwc.com/structure for further details. This content is for general information purposes only, and should not be used as a substitute for consultation with professional advisors. CH-13-0114




Daniel Eckert Managing Director, Emerging Technologies PricewaterhouseCoopers LLP [email protected]

Chris Wasden Global Healthcare Innovation Leader PricewaterhouseCoopers LLP [email protected]


About PwCPwC firms help organisations and individuals create the value they’re looking for. We’re a network of firms in 158 countries with more than 180,000 people who are committed to delivering quality in assurance, tax and advisory services. Tell us what matters to you and find out more by visiting us at http://www.pwc.com

http://www.pwc.com/structure

http://www.pwc.com


Wrapping up Phase 1 of the Mobile Innovations ForecastNew data bolster the general direction of innovation over the next five years

With this article, PwC concludes Phase 1 of its Mobile Innovations Forecast, in which we have examined trends in the performance of core components of mobile devices and infrastructure. Based on new data for these components, our fundamental assessment is that the rate of performance increases for these seven enabling components of mobile innovation—memory, application processor, storage, infrastructure speed, device speed, imaging and display technology—is expected to decelerate only slightly between 2011 and 2016, relative to 2007 to 2011.

We do see a potential trouble spot with the coming introduction of ultra high definition (UHD) video. Will the massive data streams produced by UHD overwhelm the other components? We explore that issue below. On the other hand, we are enthused by the early breakthroughs demonstrated by smartphones that use contextual information to deliver new value to owners. And a major question as we move into Phase 2 of our exploration of mobile innovation is how many mobile operating systems (OS) and associated app store ecosystems will survive to relevancy by 2016?

For new readers of the Mobile Innovations Forecast, some brief scene setting before we proceed. This forecast exists within PwC’s framework for understanding various dynamics driving the broader technology sector today, a framework that suggests ways technology companies might navigate disruptions that are rich in opportunity.

Mobile innovation is one of four market forces in this framework that are redefining customer demand, expectations and business opportunity for technology companies. The others are cloud computing, social networking and the emergence of intelligent devices. Individually, each is turning the rules of the broader technology sector upside down. Collectively, they are co-mingling in ways that paint a forward-looking picture that is starkly, even radically, unlike the past.

Our coverage of the vast mobile ecosystem is an ongoing project comprising four phases. Phase 1 examined the performance improvements of existing technology components. Phase 2, launching soon, covers new capabilities being added to mobile devices. Phase 3 will review compelling new use cases. And Phase 4 will cover new business models.

One purpose of Phase 1 has been to track the recent performance history of the core platform on which mobile delivers value and to forecast its trend line. For this purpose, PwC developed its Mobile Technologies Index, comprised of a metric for each of the seven enabling components. As part of this conclusion to Phase 1 we are reporting new data acquired from our supplier, IHS. Based on this data, we are publishing an updated version of our Mobile Technologies Index and forecasts for each of its seven technologies. Using this new data we can extend our forecast from 2015 to 2016.


2 The PwC Mobile Innovations Forecast New data bolster the general direction of innovation for the next five years

In our revised forecast, the compound annual growth rate (CAGR) for the Index, 2011 through 2016, is 36 percent [Figure 1], compared to the earlier forecast of a CAGR of 41 percent for the period 2011 through 2015 [Figure 2].

The five-point reduction in the CAGR for 2011–2016 versus 2011–2015 is mainly due to the extra year. However, three components—application processor, display technology and infrastructure speed—have downward revisions of five percentage points or more in their improvement rates between the two forecast periods. [See sidebar on page 5]

0

50

100

150

200

250

300

350

400

450

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Mob

ile d

evic

e ca

pab

ilitie

s re

lativ

e to

201

1

The red line represents a performance improvement trajectory that would match Moore’s law.

Source: IHS iSuppli Mobile and Wireless Communications Service

Figure 2: Original Mobile Technologies Index

MIFMoore’s law

41% CAGR (2011–2015)

55% CAGR (2007–2011)

PwC remains optimistic that the price-performance metrics for these basic enablers of mobile innovation will continue to improve at rates sufficient to inspire and support inventors. For the most part, these components are evolving in parallel; for example, app processors meet the data crunching needs of the OS as it delivers new gesture-sensing capabilities to users.

Figure 1: Revised Mobile Technologies Index

0

100

200

300

400

500

600

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016

Mob

ile d

evic

e ca

pab

ilitie

s re

lativ

e to

201

1

36% CAGR (2011–2016)

47% CAGR (2007–2011)MIFMoore’s law

The red line represents a performance improvement trajectory that would match Moore’s law.



The OS and the ecosystems they spawn

Although the Index does not include a metric for the OS, we cannot say often enough how key its evolution will be to future use cases involving the new technologies to be examined in Phase 2. At present, however, there are a number of questions about how the OS will evolve to support these use cases, especially across disparate mobile platforms.

Historically, desktop and server OS innovation originated from individual OEMs. The mobile ecosystem has been more complicated, making OS innovation itself more chaotic and, so far, a major source of innovation. That’s because there are different types of mobile OS environments. At present, we see three models:

• Total control—a single vendor has complete and total control of closed source code, with no relicensing and a highly coherent application ecosystem driven by the vendor;

• Shared control—a community open- source approach driven by a few large vendors where variations and extensions are managed through commercial agreements that include validation testing to limit application incompatibilities;

• Hybrid—a go-it-alone approach by a large or small vendor using an open-source code base as a starting point but making changes that introduce incompatibilities, usually in support of a different business or monetisation model.

Each type has implications for the path, or vector, of innovation. Whilst we cannot accurately predict which vector will have the greatest impact on mobile innovation, we can suggest patterns of innovation that could be more likely in each model.

The total control approach offers the vendor strong brand recognition, a track record and a critical mass of

users associated with some initial groundbreaking innovation. The challenge comes later. Once a mass of users has been established, it creates a dead weight of backward compatibility requirements. This dead weight can limit innovation to capabilities consistent with the existing ecosystem. Even if the existing ecosystem was initially launched by an earlier disruptive innovation, vendors can find it difficult later to ‘break the eggs’ by introducing a new disruptive innovation to their installed base.

In the shared control model, the core common platform offers a foundation, but introduces some freedom to build new capabilities on top that suit particular user communities. However, this potential for the OS to fork in different directions prohibits against broad adoption of any significant innovation from one fork. As such, and given network effects in technology, it is unclear if the end game of shared control is, in fact, a continuation of the model or an evolution to the total control approach.

On the other hand, depending on the nature of the open-source community and licensing terms, the most successful forks can be repatriated into the core, shared platform. This model could support a larger number of innovation initiatives, not all of which would survive but those that did could be more impactful than what a single, closed-source vendor might produce.

In the hybrid approach the level of innovation is least constrained—it is effectively a blank canvas for the most disruptive capabilities. The challenge is that a new device would need an innovation so valuable that people would be willing to carry a second device or willing to give up access to their current ecosystem of apps, or tap into an entirely new market—say in regions of the world where smartphones are not yet widely adopted.


Mobile innovation in general is about seeing what everyone sees but interpreting it differently—fighting on new ground, not on the same ground. But it is also about gaining enough traction to become widely adopted. This is an old story in innovation, and one not likely to change in the next few years.

The big takeaway

As we look at the entire picture we have painted in Phase 1, certain components stand out more than others as crucial to mobile innovation in the next few years. Although we have always held that all seven for which we have metrics, plus the OS, are enablers, and that no component alone drives innovation, especially not the disruptive variety, we now understand better that some might incite innovation more than others.

Without rehashing previous articles, let’s just note that continuing advances in display technology, imaging, infrastructure speed and application processors (quad versus single-purpose strategy more than performance per se) appear more closely tied to mobile innovation bursts.

Consider sensing technologies, to be explored more fully in Phase 2. Display and imaging will be crucial to all kinds of sensors that lead to the context- aware1 smartphones we expect. The application processer must be optimised and robust enough to handle all the additional data harvested by sensors, without quickly draining the battery or burning the user’s hand. Infrastructure speeds will need to be robust enough to move all that data back and forth to the cloud because much of the contextual awareness capabilities will reside in the cloud. Finally, the OS will be the primary manager of all this stuff on and off the device going on in some sensing application.

We expect that the metrics for the enabling components singled out in the previous paragraph will improve fast enough to keep up with the sensing-based contextual awareness capability about to explode on the scene (in ways we will explore in Phase 2).

In contrast, there are other new capabilities to be explored in Phase 2 for which realising the full capability of the new technology may have to wait on one or more enablers (which reemphasises why we designate them as enablers).

One example is UHD. UHD image sensors may be moving faster to market than the ability of the surrounding enabling technologies to keep up. UHD-capable smartphones and tablets are just on the horizon, but what do OEMs mean when they say ‘capable.’ That conceivably spans a wide spectrum from simple UHD capture of short video segments on a smartphone to watching a two-hour UHD movie that resides on a tablet.

We expect the initial use cases for UHD will be relatively modest, mostly the former example, and within the boundaries of whatever the enabling components are capable of handling. As for storing and watching or streaming a feature movie in UHD on your device, that is farther off, perhaps not within our forecast period, which is now extended to 2016. [See Figure 3]

Farther down the road there are likely to be contextual awareness applications that will need or benefit from UHD image capture. Fine-grained gesture recognition such as ‘typing in the air’ with your fingers could be one example.2 The enabling components might not be up to processing, storing and transferring the heavy quantities of data involved.

One thing that is exceedingly clear is how data centric (as opposed to voice centric) mobile is and will continue to be. The mobile device remains all about communication, but now communication extends beyond the simple phone call or text message. The amount of data collected, stored, transmitted and recovered after analysis in the cloud continues to explode. And future development, such as UHD imaging, only makes that clearer.

1 By contextual awareness we mean that a mobile device understands a user’s relationships to people, places, objects and information, and is able to infer certain needs, intents and goals of the user. Armed with this knowledge, a mobile device can meet a user’s needs and wants with minimal requirement for that person to state them explicitly.

2 http://www.sci-tech-today.com/news/HP-To-Offer-Leap-Controllers-on-PCs/story.xhtml?story_id=10000B60T69O&full_skip=1

0

50

100

150

200

250

Figure 3: UHD video requirements

UHD video is a realistic future application for mobile devices, but how much, how soon and in what form? As Figure 3 shows, UHD video will require significant amounts of storage for the content to actually reside on the device itself. Other enablers such as streaming technology, compression and playback on the device will require even further development before mobile UHD reaches full fruition. UHD is likely to drive the inclusion of more storage on smartphones and tablets, and is also likely to accelerate improvements in other technologies that might turn a tablet into a full UHD viewer. But most of that will take place outside the 2016 timeframe of our mobile innovation forecast

Source: PwC estimates

0

50

100

150

200

250

2011 2012 2013 2014 2015 2016 2017 2018 2019 2020

G b

ytes

Ultra HD video storage requirements vs mobile device capability

High-end handsets

Tablets

1 hr @ 32 Gb/hr;2 hr @ 16 Gb/hr

2 hrs @ 32Gb/hr; 4 hrs @ 16 Gb/hr

4hrs @ 32 Gb/hr; or 8 hrs @ 16 Gb/hr


As explained in the Phase 1 Introduction, the Mobile Technologies Index is a broad composite of the seven enabling components that underlie the power of the mobile device to sense, analyse, store and connect information. We created the Index as a starting point for our broader examination of mobile innovation, and have spent the preceding months looking at each of the technologies, and their metrics, component by component.

Our forecast began with these components because they are key to understanding the evolutionary curve of technological innovation, and the developments that might lead to a disruptive product that transforms an entire ecosystem. Since mobile phones appeared in the 1970s, disruptive breakthroughs have been made possible, in part, by the continuous progress of these components at predictable price points.

For this quantitative analysis, we partnered with IHS, an Englewood, Colorado-based global information and analytics provider with comprehensive databases of each sector of the high technology value chain. Using IHS data and collaborating with IHS analysts, we constructed the Index and the individual metrics we have now reported in the series of Phase 1 articles.

We now offer a revised forecast, based on data from the second half of 2012, approximately a year after the initial forecast. Individual 2011 and 2012 forecasts for each component are compared in Figure 4. A more visually illustrative way of looking at their relative improvement is the spider diagram. [Figure 5]

In the original forecast and the revised version, the comparison period of 2007–2011 was a period of even faster rates of improvement than 2011–2015 or 2011–2016. This is due largely to the disruptive nature of the first Apple iPhone, which debuted in 2007. With its huge initial popularity, followed by the immense popularity of competing smartphones, production volumes of components greatly accelerated. Higher volumes drove innovation and accelerated rates of improvement in the individual price-performance curves.

The most striking change in the new data from IHS is the 5-point slowdown in the overall CAGR of the Index, from 41 percent for 2011–2015 to 36 percent for 2011–2016. This is mainly due to the additional year in the forecast period, bringing the total now to six. The 2011–2015 CAGR recalculated with the new data is only one point different—40 percent—well within any margin of error.

That said, five individual CAGRs for components are revised downward in the new forecast: memory (DRAM) down 3 percent; imaging, 4 percent; applicationprocessors, 5 percent; display,

6 percent and infrastructure speed, 9 percent. The downward corrections are based on new data that are improved estimates of shipments, average sales prices, etc. Forecasting is, after all, a predictive endeavour susceptible to revision as estimates change.

Take application processors, for example. The CAGR for 2011–2015 in the first forecast was 53 percent in Gigahertz per dollar (GHz/$). The new CAGR, based on the IHS data for 2011–2016, dropped to 48 percent. The price-performance curve tapers off in the new forecast due to a number of factors.

Figure 4: Comparison of two forecasts

Original forecast Updated forecast

Index component CAGR (2007–2011)

CAGR (2011–2015)

CAGR (2007–2011)

CAGR (2011–2015)

CAGR (2011–2016)

Application processor 43% 53% 43% 54% 48%

Device speed 75% 37% 71% 41% 37%

Memory 49% 48% 48% 49% 45%

Storage 76% 35% 62% 39% 35%

Imaging 37% 20% 29% 24% 24%

Display technology 26% 16% 17% 10% 10%

Infrastructure 77% 54% 79% 52% 45%

Full Index 55% 41% 47% 40% 36%

Processor speed

Memory

Device connectivity speed

Infrastructure speed

Display

Storage

Imaging sensor

2006–2011

2006–2016

Figure 5: Index component changes

100 200 300 600 700 800800 700 600 400 300 200 400 500500 100



More detail about the new forecast


One is that, despite a race to build and install quad-core application processors, the sales of quad-cores has slowed because the industry now recognises that quad core is not always ideal for a mobile device the way it nearly always has been for personal computers. A dual core may run certain applications just as effectively but use less power and produce less heat—two key design factors critical for smartphones and tablets.

Meanwhile, optimisation techniques are making possible special purpose cores devoted to specific tasks now handled by identical, general-purpose multicore processors. Image processing is one example. The point is, wide adoption of quad cores by OEMs would improve the price-performance faster, but since fewer are likely to be used in mobile devices than initially predicted, quad-core price-performance slows a bit, impacting the overall metric.

Display technology is another example of products evolving in ways not anticipated in our first set of data. At the time of our initial forecast, the screen size segments ended at 3 inches. With the increased adoption of larger screens for smartphones, we can track more segments—4 inches and 5 inches—each with its own pricing dynamics. This greater granularity allows us to track more accurately what is happening with the price-performance metric for display on an aggregate basis.

The new data also provide an opportunity to validate our forecast accuracy. The first version of our Index used actual market data for the years 2007–2010 but forecasts for 2011–2015. This new version uses actual market data for 2011, allowing us to compare it to the earlier 2011 forecast. Figure 6 shows this comparison.

Four of the seven metrics are within a five percent variance. The three with greater variance were all revised downward: device speed down 10 percent, imaging down 15 percent and display technology down 19 percent.

In addition to the standard explanation of more accurate data and volatile new technologies, here are reasons these three components evolved in unexpected ways:

• Device speed, in megabits per sec-ond per dollar, is slower than forecast because LTE is rolling out more slowly than we expected, specifically outside the US.

• Imaging, in megapixels per dollar, is slower because there is greater demand for higher resolution sensors in advanced smartphones, a dynamic that is presently driving average selling prices in the seller’s direction, not in the buyer’s.

• Display technology, in performance per dollar per square inch, is lower than expected because of the shift in the technologies in use is rather volatile.

Overall, considering the dynamism of any technology market when measured on a performance-per-dollar basis, which involves unanticipated changes in demand and supply in both primary and related markets, we find the forecast accuracy for 2011 to be good. With each succeeding year of data we will continue to benchmark our accuracy, both to enhance it, and to learn more about the drivers of change in the mobile ecosystem.

-25%

-20%

-15%

-10%

-5%

0

5%

10%

Figure 6: Variances in two forecasts

Display = Performance per panel price per in2

Image = Mpix/$

Device speed = Mbps/$

Application processor = GHz/Cost

Storage = MB/$

Infrastructure = Wtd Capex w/Life cycle

Memory = Gb/$


© 2013 PwC. All rights reserved. PwC refers to the PwC network and/or one or more of its member firms, each of which is a separate legal entity. Please see http://www.pwc.com/structure for further details. PM-13-0292

About PwC’s Technology Institute

The Technology Institute is PwC’s global research network that studies the business of technology and the technology of business with the purpose of creating thought leadership that offers both fact-based analysis and experience-based perspectives. Technology Institute insights and viewpoints originate from active collaboration between our professionals across the globe and their first-hand experiences working in and with the technology industry. For more information please contact Raman Chitkara, Global Technology Industry Leader.

About PwC

PwC firms help organisations and individuals create the value they’re looking for. We’re a network of firms in 158 countries with more than 180,000 people who are committed to delivering quality in assurance, tax and advisory services. Tell us what matters to you and find out more by visiting us at http://www.pwc.com/



Let’s talk

If you have any questions about the Mobile Innovations Forecast or would like to discuss any of these topics further, please reach out to us.


Mobile Technologies Index / 1


Mobile Technologies Index Display: Enabling devices to offer users more natural interaction

The displays on smartphones and tablets will become thinner, lighter, larger and sturdier with higher resolution, improved touch sensitivity and more power efficiency through 2015, our forecast period.

These improvements will enable many of the mobile innovations we expect through 2015. Supported by the operating system (OS), the display will help the device to better accommodate the user, including better visibility in direct sunlight. Additional innovations associated with entirely new capabilities in display technologies, such as flexibility and support for haptics (tactile feedback), will be covered in Phase 2 of the PwC Mobile Innovations Forecast.

For the aspects of display technology covered here in Phase 1, PwC forecasts a compound

annual growth rate (CAGR) of 16 percent as measured in performance per dollar per square inch (P/$/in2) through 2015. [Figure 1] During 2007–2011, display technology grew at a CAGR of 26 percent, a rate sparked by the initial disruption created by the Apple iPhone.

As with other components in the PwC Mobile Technologies Index, the rate of improvement in the display CAGR is slowing, in part because the technologies are maturing. Nonetheless, PwC expects display performance and screen size to continue to increase during our forecast period, while the cost per square inch of panel will continue to decrease.

Display is the most complex and subjective metric of the Index’s seven components, which are the enabling technologies for mobile innovation. It is based on the three display


Figure 1: Display technology, compound annual growth rate (CAGR)Figure 1: Display technology, compound annual growth rate (CAGR)

0%

50%

100%

150%

200%

2008 20092007 2011 20122010 2014 20152013

CAGR 26% 2007 – 2011

CAGR 16% 2011 – 2015

Per

cent

of 2

011

Ind

ex b

ase

Performance/$/in2Performance/$/in2

Source: IHS iSuppli Small & Medium Displays Intelligence Service



http://www.pwc.com/mobileinnovations

http://www.pwc.com/gx/en/technology/mobile-innovation/mobile-device-technology-components.jhtml



technologies currently most in use, and our analysis of their cost and their performance, which is a weighted aggregation of resolution, brightness and other traits. The metric does not include the touch module, but touch technology is examined briefly in this article.

The three display technologies in the metric are active-matrix organic light-emitting diode (AMOLED) displays and two liquid crystal display (LCD) technologies—amorphous silicon (a-Si) and low-temperature polysilicon (LTPS).

Before smartphones, a-Si was the predominant display technology in handsets. As the more mature and lower performing of the three, a-Si will retain its cost advantage over LTPS and AMOLED, but will prevail only in low-end handsets.

In the high-end device market, the display already accounts for one third of the bill of materials due to the use of the more expensive LTPS and AMOLED. Because mid- to high-end smartphones are the fastest growing segments of the market, the use of LTPS and AMOLED will grow faster than a-Si. [Figure 2]

Figure 2: Handset display forecast per display type (in thousands)

Display technology Y2011 Y2012 Y2013 Y2014 Y2015

a-Si 1,785,790 1,792,150 1,800,817 1,837,979 1,884,358

LTPS 366,899 416,597 506,801 576,680 608,389

AMOLED 82,027 148,243 188,028 233,714 290,000

Grand total 2,234,716 2,356,990 2,495,646 2,648,373 2,782,747

LTPS+AMOLED 448,926 564,840 694,829 810,394 898,389

Percent of total 20.1% 24.0% 27.8% 30.6% 32.3%

a-Si – Amorphous silicon LTPS – Low-temperature polysilicon AMOLED – Active-matrix organic light-emitting diode

Source: IHS iSuppli Mobile Handset Displays Market Tracker

LTPS (used on the iPhone, for example) has the higher resolution and is more widely used than AMOLED (used on the Samsung Galaxy S III, for example). Samsung manufactures about 90 percent of AMOLED.

Our forecast captures a variety of attributes. Each display technology is following its own cost curve, so the cost factor of the metric is the average cost per square inch of the three, each weighted for its share of total square inches produced. The performance part of the metric is based on five measures and one estimate supplied by iSuppli experts: resolution, brightness, color, contrast ratio, viewing angle and power efficiency (the estimate). [See sidebar “Measuring display performance.”]

The performance equation in our metric includes contrast, but not the possibility of a game-changing innovation in contrast, especially the ability to view the display in bright sunlight. An OEM recently began shipping a device in Asia that dramatically increases the contrast ratio, which is the key to reading displays in direct sunlight.1 (To be explored in future articles)

As noted, the metric does not include the touch modules that lie just beneath the outer glass. They are typically priced and sold separately from the display technology. We expect touch technology to evolve during the forecast period.

1 http://www.techradar.com/us/news/phone-and-communications/mobile-phones/ how-future-phone-screens-will-be-viewable-in-the-brightest-sunlight-1105655

Measuring display performance Display is the only component in the Mobile Technologies Index that does not have a relatively simple performance measure. That’s because display ‘performance’ is based on a combination of several traits. The performance display metric is an aggregate weighted average of five measured traits and an estimate for a sixth trait.

The measured traits are resolution (in pixels); brightness (in luminance); color (percent of color gamut); contrast ratio (ratio of luminance between white and black— higher is better) and viewing angle (180 degrees the maximum and theoretical best).

Performance data for these traits were collected for numerous displays from various leading manufacturers that were shipping into the mobile handset LCD and OLED displays market in the fourth quarter of 2011. That data was used to generate a performance score for each of the three display technologies. We then estimated how the performance metrics would trend for each technology from 2011 to 2015. The ratings used in Figure 4 are averages based on the specifications for various screen panels provided by the manufacturers.

As for power efficiency, we estimate a 30 percent savings on average for typical usage of AMOLED (because it has no backlight) compared to the two LCD technologies.

Figure 4: Performance ratings for three display technologies

Res

olut

ion

Pow

er

Brig

htne

ss

Con

tras

t R

atio

Col

or

Vie

win

g A

ngle

a–Si 219042 1.00 300 409 0.60 135

LTPS 302947 1.00 425 597 0.65 160

AMOLED 216192 1.30 230 2500 1.00 178

Source: IHS iSuppli Small & Medium Displays Supply & Demand

Continues on page 3

http://www.techradar.com/us/news/phone-and-communications/mobile-phones/how




Handsets currently use one of two touch technologies, neither of which is especially new: resistive and projected capacitive. The latter is far more widely adopted because it enables the multi-touch capabilities of smartphones as pioneered by Apple and quickly adopted by others. [Figure 3]

Figure 3: Handset touch modules in millions of units

Application Technology 2009 2010 2011 2012 2013 2014 2015

Mobile handset

Projected capacitive

85.7 180.5 468.0 662.6 999.5 1,181.9 1,313.8

Resistive 185.2 195.2 163.1 135.4 116.8 98.0 94.9

Grand total 271 376 631 798 1,116 1,280 1,409

Source: IHS Displaybank 2012 Touch Panel Issue and Cost/Industry Analysis

OEMs are beginning to choose display glass that uses in-cell touch technology. This innovation reduces the number of layers in the handset display, making them thinner and improving touch sensitivity to the touch module underneath.

In an LCD display, the bottom layer is the backlight, followed by the LCD layer with red, green and blue pixels, and then a thin layer of glass. On top of that is the touch module, topped off by a thicker layer of glass. So the middle layer of glass separates the liquid crystal portion from the touch portion. AMOLED is similar, but does not require the backlight layer because each LED provides its own light.

Emerging in-cell touch technology eliminates the layer of glass in the middle and combines the touch with the LCD or OLED layer into a single layer. This reduces the distance between the user’s finger and the touch capability, enabling taps and gestures to be more responsive than they are now.

Touch innovation is an example of the interdependencies amongst the components in the Index. The resistive and projected capacitive touch technologies have been around for a decade or longer, but other components had to advance before today’s touch capabilities were realised. For example, the UI (user interface) needed increases in processing power, more sensitive and power-friendly displays and advances in the part of the OS that manages the UI.

We expect these interdependencies to continue, with incremental innovation, such as in-cell technology, supported by processor power and more sophisticated OSes.

As noted, the metric does not include new display technologies that might eventually compete with a-Si, LTPS and AMOLED. For example, one OEM has recently shipped the first smartphone with an indium gallium zinc oxide (IGZO) display, which has a higher pixel count, high-quality color, lower cost and lower power usage than LTPS. It also requires significantly less backlighting.2

Whether lower power use and higher resolution come from IGZO or from future versions of LTPS and AMOLED, those improvements will be crucial to enabling smartphones and tablets to display higher resolution content, such as the impending Ultra High Definition standard, which we expect to start rolling out during the forecast period.

2 http://www.electronista.com/articles/12/11/27/japan.only.handset.uses.low.power.display.technology/#ixzz2EOL4ywQa

Using a-Si as the base of 1, we calculated the absolute performance differences for LTPS and AMOLED for each trait. [Figure 5]

Figure 5: Absolute performance difference

Res

olut

ion

Pow

er

Brig

htne

ss

Con

tras

t R

atio

Col

or

Vie

win

g A

ngle

a–Si 1.00 1.00 1.00 1.00 1.00 1.00

LTPS 1.38 1.00 1.42 1.46 1.08 1.19

AMOLED 0.99 1.30 0.77 6.11 1.67 1.32


We then assigned a relative weighting for each performance trait, based on its deemed importance to overall display performance. This weighting is subjective, based on the goals and market positioning of today’s handset OEMs. Resolution is highest at 3, power efficiency is 2, brightness is 1 and the other three are one-third each. Using these weightings, we calculated the relative performance differences and a total perfor-mance score for each technology. [Figure 6]

Figure 6: Relative performance differences

Performance weighting: 3 2 1 0.33 .033 .034

Res

olut

ion

Pow

er

Brig

htne

ss

Con

tras

t R

atio

Col

or

Vie

win

g A

ngle

Tota

l Per

f S

core

a–Si 3.00 2.00 1.00 0.33 0.33 0.34 7.00

LTPS 4.15 2.00 1.42 0.48 0.36 0.40 8.81

AMOLED 2.96 2.60 0.77 2.02 0.55 0.45 9.34


Continues on page 4

http://www.electronista.com/articles/12/11/27/japan.only.handset.uses.low.power.display.technology

http://www.electronista.com/articles/12/11/27/japan.only.handset.uses.low.power.display.technology

“The recent Consumer Electronics Show [January 2013] in Las Vegas convinced many that Ultra HD is so good it is likely to drive the next replacement cycle for higher performing display technologies,” says Daniel Eckert, PwC Director for Emerging Technologies. “However, there will also be demand for low-cost displays to support the next 2 billion mobile phone subscribers who will be coming from emerging markets like Africa and Asia. These markets will demand low-cost, low-power displays that will be embedded in phones costing 10-30 Euros. These displays could use e-ink, which supports bright light and usage times of over a week, or updated low-resolution and low-power color displays that will address requirements for these emerging markets.”

Any improvement in display technology that impacts the visual clarity or the touch sensitivity is inherently useful. Higher resolution, brighter screens, better readability—especially outdoors—greater touch sensitivity and gesture sensing will make displays more useful, whilst also enabling entirely new apps not previously possible.

For example, in the application processor article we introduce a future scenario in which video content can be streamed from the cloud, through the smartphone via WiFi Direct to be displayed on a TV screen. In this use case, the smartphone becomes the remote control for the TV set. Within our forecast period we expect new touch-free technology (to be explored in a future article) that senses where your fingers are hovering over the device screen and projects an icon on the TV screen for the button you are above—channel changer or volume control, for instance.

Display technologies of the not-too-distant future will enable higher end devices to require less of humans and to offer more intuitive, natural qualities to better accommodate the user. The most potentially disruptive of these may be the wearable computing trend, where the traditional display is replaced (in part) by an image embedded in a pair of glasses. This broader trend toward less disruptive, more ‘heads up’ display technology will be covered in Phase 2 of our research.


Raman Chitkara Daniel Eckert Global Technology Industry Leader Emerging Technologies Director PricewaterhouseCoopers LLP PricewaterhouseCoopers LLP [email protected] [email protected]


About PwCPwC firms help organisations and individuals create the value they’re looking for. We’re a network of firms in 158 countries with more than 180,000 people who are committed to delivering quality in assurance, tax and advisory services. Tell us what matters to you and find out more by visiting us at http://www.pwc.com/

Then using a-Si as the base again, we con-verted the total performance score into a relative performance score (relative to a-Si) for 2011. [Figure 7]

Figure 7: Relative performance score

2011

a–Si 1.00

LTPS 1.26

AMOLED 1.33


We multiplied the relative performance scores for each technology by the number of hand-sets for each technology, and summed those to get the total score for all types. This total score was divided by the total number of handsets to get the volume weighted average performance score per handset for the metric. [Figure 8]

Figure 8: Volume weighted average performance score

Handset displays shipped

Display technology Y2011

a–Si 1,785.8

LTPS 366.9

AMOLED 82.0

In millions Grand total

2,234.7

Weighting of relative performance per display Y2011

a–Si 1.0

LTPS 1.26

AMOLED 1.33

[See sidebar on page 3]

Total score per display type Y2011

a–Si 1,785.8

LTPS 461.7

AMOLED 109.5

Total ‘score’ of all display types combined

2,356.9

Total handset displays 2,234.7

In millions Volume weighted average performance ‘score’ per handset

1.05


Daniel Eckert

Emerging Technologies Director

©2013 PwC. All rights reserved. PwC refers to the PwC network and/or one or more of its member firms, each of which is a separate legal entity. Please see http://www.pwc.com/structure for further details. This content is for general information purposes only, and should not be used as a substitute for consultation with professional advisors. BS-13-0219.0213

http://www.pwc.com/gx/en/technology/mobile-innovation/application-processors-enable-device-innovation.jhtml

mailto:[email protected]

http://www.pwc.com



Mobile Technologies Index Image sensor: Steady growth for new capabilities

Introduced in 2000, camera phones—combined with social media—continue to redefine what it means to communicate. The early camera phones produced low-quality images, but now, thanks to Moore’s Law, they are competing with point-and-shoot cameras and camcorders. How much more can camera phones do?

Quite a bit more, it turns out. The innovation curve will continue to produce camera phones that capture higher quality still images and videos with better sound. These improvements will result from image sensor improvements and increasingly from technology around the sensor, especially the use of microelectronic mechanical systems (MEMS), as discussed below.

Over the longer term, compelling new use cases that rely on tighter integration between the camera module and the operating system will lead to new capabilities more aptly classified as

machine vision. We introduce this topic below, but anticipate a deeper analysis in a future article. The main focus of this article is the image sensor, which is a component of the PwC Mobile Technology Index. The Index comprises seven technologies that enable mobile innovation.

PwC forecasts a compound annual growth rate (CAGR) of 20 percent for image sensors as measured in megapixels per dollar (MP/$) through 2015. [See Figure 1] Since 2007, image sensors have followed Moore’s Law, doubling megapixel density per dollar every two years for a CAGR of 37 percent. The MP/$ will continue to grow, but at a slower rate. [See Figure 2] This, however, will not reduce or depreciate the image sensor’s importance to mobile innovation. To understand the future of image capture, it is useful to review the evolution of imaging in smartphones.

Figure 1: Image sensor, compound annual growth rate (CAGR)

Per

cent

of 2

011

Ind

ex b

ase


20% CAGR (2011–2015)

37% CAGR (2007–2011)

2007 2008 2009 2010 2011 2012 2013 2014 20150%

50%

100%

150%

200%

250%MP/$



2 Mobile Technologies Index Image sensor: Steady growth for new capabilities

Figure 2: Price performance compared to Moore’s Law

Meg

ap

ixel p

er

$


2007 2008 2009 2010 2011 2012 2013 2014 2015

0.50

1.15

1.80

2.45

3.10

3.75

4.40

5.05

5.70

6.35

7.00

CCD CMOS Moore’s law

The image sensor is part of a camera module for the mobile device that also includes the lens and MEMS for various (and a growing number of) functions, plus interfaces to the application processor, memory and storage.

The first camera modules for cell phones used charge-coupled device (CCD) sensors. As recently as 2006, 4MP CCD-based modules cost US$22, which included US$8 for the image sensor. Camera module suppliers charged as much as US$15 more for ‘integration’ of this subsystem into the handset. Aside from being costly, the early camera modules came in only one size (large) and used significant amounts of power, consuming the battery faster.

The lower quality images from early camera phones were relegated to viewing on the phone screen because the resolution was not high enough for printing or displaying on a computer screen. As well, sharing photos on 2G/2.5G broadband cellular networks was much slower, and most camera phones didn’t yet support WiFi or Universal Serial Bus (USB) standards.

As a result, there were many hurdles to sharing, posting and permanently saving images. Despite these obstacles, camera

phones served a practical purpose and became a standard feature in 80 percent of all handsets manufactured—increasing to 90 percent within our forecast period.

CCD-based handsets have been slowly phasing out since the introduction of complementary metal-oxide semiconductor (CMOS) sensors in 2006. CMOS sensors consume less power than CCD sensors. They also enable lower cost, smaller form factors and much higher image quality.

Consumers now expect to capture higher-quality photos and to more easily share them in email, on social networks and as prints. They can look forward to the quality of images and video to continue to improve year after year. By 2015, high- end smartphones, which already have 8MP image sensors, will enter the 14MP to 20MP range.

Higher resolution image sensors equal to or approaching those of the point-and-shoot digital camera are only one part of the formula for image improvement. Through software manipulation and various MEMS devices, smartphones will continue to deliver improvements in image stabilisation; auto focus; zoom; light sensitivity; low-light performance; noise reduction; reduced power use;


integrated image processing hardware; sound recording for video and lens improvement. Software will also become a differentiator in image collection, compression, picture editing, video editing and searching.

MEMs are crucial. They are made with standard semiconductor manufacturing processes and benefit from the predictable price-performance curve of those processes.

Software and MEMs have already enabled new or improved functions in some smartphones: autofocus1 (MEMs), panoramas from multiple photos (software) and high-definition image capture (MEMs and software).

Improvements in image sensors, MEMS and software will enable future-generation camera phones to offer 3D imaging. 3D will not only provide a richer, more realistic photographic image, but the 3D depth map will be crucial to the sensing required for hands-free gesture control of the device and improved facial recognition.

1 http://www.doc.com/Actuator/Pages/Actuator.aspx

Due to these continuing improvements in mobile device imaging—still and video—the market for point-and-shoot digital cameras is likely to contract. In 2011, the average handset had 30 percent of the MP count of the average digital still camera. By 2015, handsets will have 60 percent of the MPs in digital still cameras on average, with some having more than 80 percent. [See Figure 3]

“In the near future, smartphones will have more than enough raw pixel resolution and capabilities to meet most users’ needs, and those users will become less willing to pay a premium for additional megapixels, unless new capabilities, like machine vision, are added,” says Robert A. Chinn, a principal in the Semiconductor Advisory Practice at PricewaterhouseCoopers LLP.

Figure 3: Megapixels in handsets vs. digital still cameras (DSC)

Meg

ap

ixel


2010 2011 2012 2013 2014 20150

4

8

12

16

20

Avg handset Leading-edge handset Digital still cameras

Avg handset MP @ 60% of DSC

Avg handset MP @ 30% of DSC

Robert A. Chinn

Semiconductor Advisory Principal PricewaterhouseCoopers LLP


As noted earlier, improved image sensors and camera modules will factor into machine vision capabilities that are just now appearing on the horizon. Machine vision broadly comprises the translation of light (visible to the human eye or not) into digital information, and the analysis of that digital data for the purposes of identifying and extracting information about objects of interest. To date, machine vision has mainly been associated with manufacturing for quality control and other industrial uses. In smartphones, machine vision will involve the use of the image sensor to capture information that is analysed in the device or in the cloud (or both), and put to some personal purpose. For example, there already is a smartphone app that uses the image sensor to help it detect and measure gamma rays—radiation—in the immediate area.2

Machine vision capabilities in mobile devices will use the image sensor in different ways than photo capture does. Some machine vision applications will use less information than required for photos; for example, make determinations based on examining the contours of objects in an image but without needing a full color rendering of it.

Machine vision use cases supported by all the anticipated advances in image sensors, MEMs and software could include hands-free gesture control, facial

2 http://www.vision-systems.com/articles/2012/05/smartphones-measure-radioactivity.html

recognition and mobile industrial and medical applications. Another recently announced non-photo use of the image sensor is an infrared (IR) sensor that takes the body’s temperature. The IR sensor, which is a MEMs device, is designed to sit next to the image sensor in the smartphone and use its viewfinder to target the correct spot of the object or subject to take a temperature. We will explore machine vision use cases like this in depth in future articles.

Because the camera modules will move from commodity parts with limited functionality to multi-purpose modules capable of wide ranging uses, the modules will require deeper integration with the operating system. PwC anticipates consolidation in the vendor space within our forecast period as fewer suppliers will be likely to meet the higher threshold for functionality and OEMs will look to establish long-term relationships with individual suppliers. In 2012, the supply base included 40 image sensor suppliers and 40 camera module suppliers.

Whatever new use cases image sensors end up supporting, camera modules will clearly evolve within our forecast period to support everything needed for higher quality video and still photography, and use cases associated with recording scenes for later viewing or playback.



About PwC

PwC firms help organisations and individuals create the value they’re looking for. We’re a network of firms in 158 countries with more than 180,000 people who are committed to delivering quality in assurance, tax and advisory services. Tell us what matters to you and find out more by visiting us at http://www.pwc.com/.


© 2013 PwC. All rights reserved. PwC refers to the PwC network and/or one or more of its member firms, each of which is a separate legal entity. Please see http://www.pwc.com/structure for further details. BS-13-0199


Robert A. Chinn Semiconductor Advisory Principal PricewaterhouseCoopers LLP [email protected]

Let’s talk

If you have any questions about the Mobile Innovations Forecast, or would like to discuss any of these topics further, please reach out to us.



Mobile Technologies Index Storage: Quenching the thirst for more

As smartphones and tablets take on more of the computing chores that laptop and desktop com-puters used to handle, one might wonder if stor-age capacities will keep pace with the growth in digital content captured, shared and stored on these devices. Users won’t be disappointed.

Given the continuing improvement of the price-performance curve for solid-state flash memory, we predict robust growth in the amount of storage in mobile devices. For reasons we will explain, this is so even as cloud-based storage grows.

Specifically, PwC forecasts a compound annual growth rate (CAGR) of 35 percent for NAND flash memory, as measured in Megabytes per dollar (MB/$), through 2015. [Figure 1] The CAGR for flash memory is fourth amongst the seven enabling technologies of the PwC Mobile Technologies Index, our key indicators of mobile innovation trends.

NAND is the storage component in our Index because it is used in solid-state drives (SSD). Mobile devices created an early market for SSDs, replacing larger mechanical hard drives. Now, mobile innovation is benefiting from the growing demand for solid state in all other computing devices, and the volume-driven, price-performance improvements the broader demand sustains.

“Solid state is now front and centre for the entire computing industry,” says Robert A. Chinn, a principal in PwC’s Semiconductor Advisory Practice. “While solid-state drives are replac-ing the spinning hard disk drives on PCs, the higher cost of the drives has been a barrier in their widespread adoption in PCs. A continuing decline in cost of solid-state drives will further reduce costs for mobile devices.”

Robert A. Chinn

Semiconductor Advisory Principal

Figure 1: NAND flash memory, compound annual growth rate (CAGR)

Per

cent

of 2

011

Ind

ex b

ase


35% CAGR (2011–2015)

76% CAGR (2007–2011)

2006 2007 2008 2009 2010 2011 2012 2013 2014 20150

50

100

150

200

250

300

350MB/$




2 Mobile Technologies Index Storage: Quenching the thirst for more

Our 35 percent CAGR forecast means that OEMs in 2015 will be able to install more than three times as much (3.3x) NAND for the same price as they do today. Average flash capacity in high-end handsets is already 22GB, and will grow to 50GB by 2015. [Figure 2] These numbers are averages—some smartphones and tablets already offer 64GB. We expect to see some 128GB tablets soon and some 256GB tablets are likely next year.

We use NAND flash, not NOR. They both have the main defining characteristic of flash memory—they are non-volatile. But NAND supports much higher capacities, and is used to store music, video, photos, contacts, emails and other data in SSDs, USB drives and memory cards. If a mobile device contains NOR flash, it is typically used for operating system purposes.

As noted, SSDs were first used in mobile devices. Specifically, Apple replaced the mini-disk drive with an SSD in the iPod as a proving ground for what it would do with the first iPhone. The new, thinner form factor of the SSD has meant that solid state has become the standard for smartphones and tablets. As the price curve dropped, OEMs also began to use solid state in laptops, desktops and serv-ers. SSDs are the fastest growing segment of storage drives [Figure 3], but are not likely to overtake HDDs any time soon. [Figure 4] Mobile HDDs will also continue to be used in some notebooks, laptops, game boxes and other devices.

Figure 2: Average storage in devices in GB

2008 2009 2010 2011 2012 2013 2014 2015

Handsets 1 2 3 4 6 8 10 12

High-end handsets 5 10 12 17 22 29 38 50

Tablets 29 35 46 61 75 92


Source: Gartner

HDD SSD Mobile HDD

2012 2013 2014 2015 20160%

20%

40%

60%

80%

100%

120%

140%

Figure 3: Total storage capacity comparisonYear-over-year growth


What will users do with three-times the storage they have now? Much of it is likely to be used to store more HD video and photos at higher levels of resolution. A 128GB drive could store HD versions of the last four or five Oscar-winning films on your tablet. [Figure 5]

Image and video capture capabilities are improving (examined in a future article), their resolutions (megapixels)

are increasing and the ability and desire of users to share them on social networks is accelerating. Shared content among personal networks will consume more storage. Television ads are already touting the ability to swap music play-lists and personal videos between two smartphones by simply placing them next to each other. We will explore these use cases in future articles.

Figure 4: Total storage drive capacity to be producedHard drive and solid-state drive total capacity

Pet

abyt

es

Source: Gartner

HDD SSD

2012 2013 2014 2015 20160

500

1000

1500

2000

2500

3000

Figure 5: Media capacities

A typical mix of video in hours MP3s and photos in thousands of units

Source: PwC estimates

32GB 64GB 128GB 256GB

HD 1,080p Hrs

HD 720p Hrs

MP3 K Units

Photos 16MP K Units

Photos 3MP K Units

0

20

40

60

80

100

120

140

32GB 64GB 128GB 256GB0

20

40

60

80

100

120

140

8 8 7 1116 13

22

3227

45

64

54

90

16 16

32 32

64 64

128


Cloud-based applications and storage ser-vices, such as those from Amazon, Apple, Google and others, will also factor into mobile use cases and business models. Our analysis of the cloud’s impact on flash memory might seem counterintuitive: we anticipate that cloud-based storage ser-vices will actually increase the demand for on-device storage. We expect the two to grow together over the five-year forecast period, although we do not track cloud storage in our Index.

The weakness in the argument that most content will be stored in the cloud is that it assumes sending data back and forth to the cloud will be fast and affordable. As the thirst for more data over broadband grows, it will risk clogging the network and/or driving up personal usage bills. Connecting every time, any time, in real time to the cloud over cellular networks to upload gigabytes of HD video or stream HD movies will simply not be feasible because broadband connectivity will con-tinue to vary widely in cost, speed, latency and ubiquity.

Users already determine which content is best suited to the cloud and which is best suited to the device. They understand the cost of moving very large quantities of content back and forth over cellular broadband, and make use of Wi-Fi net-works to download or update content on devices. They will continue to consume a lot of content from local storage on the device without interruption and without incurring connectivity costs.

In the processor article, we noted a new use case that requires more processor power: the ability to stream content wirelessly from a smartphone or tablet directly to a TV set or computer screen. The content could stream from the native storage on the device or from the cloud through the device. We expect many users will download content to the device and use it from there.

Then too, there are new standards on the horizon for ultra-high definition television (UHDTV). Either one of the proposed digital formats, 4K and 8K, would challenge storage and cellular broadband capacities as we now know them. The compression standards are still developing, but we anticipate storage requirements for 4K and 8K video will be more than four and eight times, respectively, greater than today’s HD movies, based on pixel resolution alone. Other traits, such as higher color density and frame rate options as high as 120 frames per second, mean the storage requirements could be much larger.

With enough embedded storage—some-day beyond our forecast period a tera-byte is not inconceivable—a user could store a standard HD version of every John Wayne movie or the entirety of “The Sopranos” TV series, but if there is adequate public WiFi access at a reason-able price, a mobile user might choose to access the same content on a streaming basis from Amazon or Netflix.

http://www.pwc.com/gx/en/technology/mobile-innovation/application-processors-enable-device-innovation.jhtml


The Technology Institute is PwC’s global research network that studies the business of technology and the technology of busi-ness with the purpose of creating thought leadership that offers both fact-based analysis and experience-based perspectives. Technology Institute insights and viewpoints originate from active collaboration between our professionals across the globe and their first-hand experiences working in and with the tech-nology industry. For more information please contact Raman Chitkara, Global Technology Industry Leader.

About PwC

PwC firms help organisations and individuals create the value they’re looking for. We’re a network of firms in 158 countries with close to 169,000 people who are committed to delivering quality in assurance, tax and advisory services. Tell us what matters to you and find out more by visiting us at http://www.pwc.com/.





Let’s talk

If you have any questions about the Mobile Innovations Forecast, or would like to discuss any of these topics further, please reach out to us.

At this stage, there are more questions than answers about the relative use of on-board stor-age versus cloud storage. We will explore these issues in detail when we examine use cases and business models.

One thing is clear right now. Friends like to share photos and videos with friends. There are already options that allow sharing amongst friends automatically through synchronisa-tion services. An individual shoots a video with his smartphone and friends automati-cally see it filed in their local storage on their mobile devices.

Thus, local storage in the form of NAND flash memory will be a key enabler of what we expect will be a series of new behaviours from a different kind of accelerated sharing on tomorrow’s social networks to the ability to consume all manner of content on the mobile device, whether it is connected or not.



Mobile Technologies Index Memory: The ever-predictable DRAM path

In a world constantly unsettled by disruptive innovation, it is comforting to know there are a few things you can count on. The phenomenal price-performance curve for Dynamic Random Access Memory (DRAM) has been a predictable constant for more than 30 years. It is easy to take this for granted.

We believe the long-term trend will continue. We forecast a compound annual growth rate (CAGR) of 48 percent for DRAM as measured in Gigabits per dollar (Gb/$) through 2015 (see Figure 1), compared to a 49 percent CAGR in 2007-2011. This means the growth rate of improvement in memory will be second only to the expected gains in processor speed (53 per-cent CAGR in GigaHertz per dollar) among the seven components of the Mobile Technologies Index, which are our key indicators of mobile innovation trends.

“The continued dramatic increases in the amount of affordable memory will lead directly to growth in the capabilities and innova-tions of operating systems, applications and use cases that mobile devices will be able to handle during the five years of the PwC Mobile Innovations Forecast,” predicts Daniel Eckert, PwC Director for Mobile Computing.

While the CAGR as illustrated in Figure 1 con-tinues a decades long trend, what is changing is that we have now reached a critical level of performance at which mobile devices can meet and exceed (in combination with other capabili-ties) the wide range of uses previously seen only on laptops and desktop computers.

Our analysis of 30-plus years of data shows a 55 percent CAGR in Gb/$ since 1980, meaning our forecast is consistent with history. This does not mean that DRAM is immune to the semiconduc-tor industry’s boom and bust cycles, as the past decade illustrates.

First, consider that from 2004 through 2008, chip manufacturers achieved CAGR improve-ments of 70 percent per year. This was due to three factors:

• an increase in capital spending by chip mak-ers after the dot-com implosion and before the 2008 recession, especially 2003 to 2006;

• although the transition began earlier, 12-inch wafer capacity surpassed 8-inch capacity during 2004-2008, dramatically increasing the available volumes;

• consistent growth in demand, due to the surge in purchases of laptops, netbooks and other mobile devices.

Daniel Eckert

PwC Director for Mobile Computing


2 Mobile Technologies Index Memory: The ever-predictable DRAM path

Next, note that during the global reces-sion (2008-2010), the growth of DRAM in Gb/$ slowed to 20 percent per year. Overcapacity also resulted in price de-clines, leading some manufacturers to close some fabs. This has been factored into our forecast.

Looking ahead, we anticipate 48 percent CAGR in DRAM based on several factors: renewed capital expenditure that began in 2011; improving worldwide gross domestic product (GDP) relative to the recession and the rapid growth in markets for smartphones and tablets. Continued sales of PCs of all types will also contrib-ute to the continued price-performance improvements in DRAM.

Computing power, memory capacity and storage (the topic of our next article) all work together to create the user experi-ence. As the amounts of each increase in tablets and smartphones relative to tra-ditional PCs, mobile devices will become capable of performing functions and run-ning applications previously associated only with PCs.

By 2015 average smartphones will have 40 percent of the 4GB of DRAM that PCs on average have today. Likewise, by 2015 the top 10 percent of smartphones will have 65 percent of the DRAM that PCs have today, and will perform correspond-ingly. The trend for tablets is even more dramatic (see Figure 2). Figure 3 shows the actual amounts of DRAM in gigabytes, on average, over our forecast period.

Already on the market, the iPhone 5 and the Samsung Galaxy S III both have 1GB of DRAM. The iPad 3 also has 1GB of DRAM, while the Samsung Galaxy Note II, a hybrid smartphone-tablet, has 2 GB.

DRAM represents 5 percent to 15 percent of a smartphone’s bill of materials. That portion is likely to remain steady over the five-year forecast because OEMs will load more DRAM on their devices as the price–performance curve improves, and will offer more applications that need it.

Figure 2: Mobile device DRAM as percentage of today’s average PC DRAM

2010 2011 2012 2013 2014 2015

Smartphones 6.6% 10.5% 16.4% 23.8% 31.8% 41.3%

Top smartphones 10.1% 14.7% 21.3% 30.9% 44.8% 65.0%

Tablets 6.3% 15.6% 29.7% 43.5% 61.9% 85.3%

Source: IHS iSuppli Memory and Storage Service

Figure 1: DRAM, compound annual growth rate (CAGR)

Per

cent

of 2

011

Ind

ex B

ase

Source: IHS iSuppli Memory and Storage Service

48% CAGR (2011–2015)

49% CAGR (2007–2011)

2006 2007 2008 2009 2010 2011 2012 2013 2014 20150

100

200

300

400

500Gb/$

3 Mobile Technologies Index Memory: The ever-predictable DRAM path

Figure 3: Mobile device DRAM averages

Gb

/Uni

t D

RA

MSource: IHS iSuppli Memory and Storage Service

Smartphones Top smartphones Tablets PCs

2010 2011 2012 2013 2014 20150

2

4

6

8

10

.21

.32

.20

3.20

.41 .58

.62

3.95

.78 1.01 1.

41

4.73

1.35 1.

76 2.47

5.69

2.24

3.16

4.36

7.04

3.62

5.70

7.48

8.76

DRAM is a key enabler of mobile inno-vation. DRAM and the central process-ing unit are the heart and soul of any computing device. And today’s mobile devices are powerful computing machines because of the amount of DRAM and powerful processors. But simply scaling up DRAM on the motherboard doesn’t address the speed at which processors can consume data. Higher functionality often means placing more memory on the processor itself.

“Performance related to DRAM includes more than just increased density and total Gb/$. Chip and system designers continue to look for ways to get past the ‘memory wall’ caused by the latency and limited bandwidth of traditional off-processor memory,” says Robert A. Chinn, a Principal in PwC’s Semiconductor Advisory Practice. “Companies are explor-ing new technologies and both existing and new architectures to more tightly couple memory with CPUs/GPUs/APUs to improve overall performance.”

In later reports of the Mobile Innovations Forecast, we will examine future use cases in more detail, however, for purposes

of positioning DRAM in the Mobile Technologies Index, consider how DRAM has enabled today’s mobile capabilities.

Smartphones and tablets would not exist as we know them if DRAM price–perfor-mance improvement had not increased at such a dramatic rate. DRAM is the secret sauce behind many mobile capabilities that we take for granted today:

• Camera functionality from digital stills to HD video capture and processing;

• Enhanced displays from increasing screen size, to HD level resolution to emerging 3D capabilities and multi-touch support;

• Running multiple applications and being able to seamlessly switch between them.

Smartphone functions that are now standard, such as Bluetooth, Wi-Fi, GPS and motion sensing, would not have been possible without the availability of lower cost, denser DRAM. These functions are enabled largely by advances in smart-phone operating systems. And operating systems, such as iOS, Android, Blackberry OS, Windows Phone and others, need ever increasing amounts of DRAM to func-tion. Today’s operating systems were not feasible in a hand-held device even five years ago because the cost of providing the necessary DRAM would have pushed device price points beyond acceptable levels to consumers.

Robert A. Chinn

Principal in PwC’s Semiconductor Advisory Practice




Daniel Eckert Mobile Computing Director PricewaterhouseCoopers LLP [email protected]


Let’s talk


During the five-year forecast period, continued improvement in its CAGR will allow greater amounts of DRAM on mobile devices to sup-port new use cases that involve more data, more computing power and more and different multi-processing.

The drivers of new use cases will tend to be processor speed, network speed, image process-ing and software, not DRAM. But by becoming more affordable, more DRAM on the device will not constrain these innovations, and will in fact support them when they become feasible. More affordable DRAM will support use cases like these:

• The bit rates from a new HD standard, called ultra-high definition television (UHDTV). There are two proposed digital formats: 4K UHDTV with 8.3 megapixel (MP) resolution and 8K UHDTV with 33.2 MP resolution (16 times the number of pixels in the current HDTV standard);

• Multiple HD video streams on a tablet for a telepresence group experience;

• Highly immersive 3D gaming experiences including precise gesture control of game mechanics and

• Simultaneous application processing with high DRAM requirements.

The latter use case anticipates an evolution from today’s multi-processing scenarios, which position one application as dominant and background applications waiting to be switched into the foreground for user interaction. New use cases will emerge in which background applications are actively processing environ-mental inputs without direct user interaction, even as a foreground application maintains user attention.

Consider the following future scenario: You are at a business conference running an app that captures the voice of the speaker, converts the speech to text and automatically summarises the content. Now you need to respond to a com-plicated incoming email from your office. The device continues to run the content capture and summarisation app even as you are answering your email. While all this was going on, your device also scans the room for colleagues enter-ing the auditorium, and lets them know you are here, too, via a text message.

This kind of scenario, with multiple applica-tions simultaneously processing large amounts of information, will drive the specification for more DRAM in future mobile devices.

In summary, more affordable DRAM will continue to be a major supporter of mobile innovation.


About PwCPwC firms help organisations and individuals create the value they’re looking for. We’re a network of firms in 158 countries with close to 169,000 people who are committed to delivering quality in assurance, tax and advisory services. Tell us what matters to you and find out more by visiting us at http://www.pwc.com/.



Mobile Technologies Index Application processors: Driving the next wave of innovation

Application processor speed is improving so fast that by the time you finish reading this article, it is likely to have made another leap forward and triggered a new wave of mobile innovation.

While that’s an exaggeration, it is not an exaggeration to expect application processors’ clock speeds to increase from 1Gigahertz (GHz) to 1.5GHz and beyond in 2012. This is important because a jump of this magnitude has historically marked the beginning of the next cycle of processor-driven innovation for mobile devices, as explored later in this article and illustrated in Figure 2.

PwC forecasts a compound annual growth rate (CAGR) of 53 percent for application processor speed, as measured by Gigahertz per core per dollar (GHz/Core/$), through 2015 [see Figure 1], a faster growth rate than the 43 percent CAGR for 2007-2011. Stated another way, by the end of our forecast period, application processor speed will have increased five times from what it was in 2011, our baseline year. The CAGR for application processors will grow faster than the CAGR for any other device component of the Mobile Technologies Index, and second only to the infrastructure speed component (54 percent CAGR).

This essentially means that the application processor, which is equivalent to a personal computer’s central processing unit (CPU) and is also known as the mobile processor, will enable various mobile device use cases,

likely to include more powerful multitasking operating systems, more immersive and natural user interfaces and more powerful graphics, including 3D. Many vendors are already designing and building these because they can be sure that the processor will be up to the tasks.

We define processing power as an amalgam of the maximum clock rate per core with which a mobile device can perform tasks and run applications. In smartphones and other mobile devices, application processors have evolved from single core 300-400MHz chips to dual core 1.2 to 1.5GHz chips and are on the road to quad core 1.5GHz and beyond. Several tablets already have 1.66GHz processors.

To understand the importance of the application processor to mobile innovation, consider the user interface. The original 2G iPhone, launched in 2007, was the first smart phone of its type to allow for a multi-touch user interface. And this was around the same time that the GHz/Core/$ metric allowed application processor chips to surpass the 500MHz threshold.

The earlier application processors used by smart phones were more commonly in the 200 MHz to 400MHz range and limited operating systems to point and click interfaces using trackballs, wheels or styli as well as single-touch capability.


2 Mobile Technologies Index Application processors: Driving the next wave of innovation

The advent of multi-touch spawned a wave of innovation aimed at more natural user behaviour when flipping and viewing pictures and pages or for other tasks. The user interface of a smart phone is now more similar to the hand gesture user behaviour for the same tasks in the physical world—with all the “intuitiveness” that implies—and a big driver of the mass adoption of these devices.

Another example later in the smart phone evolution is that of multi-tasking operating systems. As the smart phone evolved from a personal information device into a full-fledged mobile computing platform, consumers expected to be able to work on and share data between multiple applications concurrently as they do on a PC, pulling operating system requirements in that direction.

However, it wasn’t until 2009-2010 when the GHz/Core/$ metric enabled mobile devices to reach the 1GHz threshold that this type of operating system became viable and able to be fielded in such a way that did not bog down the device to the point of rendering it unusable.

Limitations to overcomeAlthough mobile processing power has begun to approach personal computer processing power, smart phones and tablets have constraints that suggest full equivalence with the PC may be well beyond our forecast period. As gigahertz increase in the smart phone, two problems arise: power limitations and overheating.

Since we only expect incremental improvements in battery life and power management during our forecast period, something will need to be done to allow multi-core processors to operate without quickly draining the battery. And given the tiny footprint of the processor in handheld devices, plus the lack of a fan, the heat dissipation problem could become acute. A user won’t get much use from all the cores if the device itself is too hot to hold.

The industry move to the 28 nanometer silicon process node in chip manufacturing could help, but chipmakers are also working on other solutions. In standard multicore architectures as seen on desktop computers, the processing for any job is typically spread among undifferentiated cores, which heats up everything and is not always the most efficient use of power.

In mobile, one approach could be “smart multicore,” in which individual cores are specialised for a particular processing requirement—say video—thereby reducing power drain and heat generation. In other words, cores could be power-optimised for the specific job. A variation on this is a separate adjunct chip for a specific computing task. In other words, in the context of future mobile devices all gigahertz may not be created equal. We will track this evolution, and depending how it goes, we might need to rethink the composition of the application processor metric in the Mobile Technology Index.

Figure 1: Applications processor, compound annual growth rate (CAGR)

Per

cent

of 2

011

Ind

ex b

ase


53% CAGR (2011–2015)

43% CAGR (2007–2011)

2006 2007 2008 2009 2010 2011 2012 2013 2014 20150%

100%

200%

300%

400%

500%

600%GHz/Core/$

3 Mobile Technologies Index Application processors: Driving the next wave of innovation

Figure 2: Cycles of processor-driven innovation

Gig

aher

tz


The industry will continue to push beyond the 1.66GHz threshold in application processors in 2013. But how far beyond depends on factors not currently predictable. The possibility of 3GHz exists—PC processors have gone there. But power and heat problems for mobile devices at that clock speed are creating huge challenges for chip designers. Moore’s Law is silent on these issues.

Multitouch

Multitasking

Next-genimmersiveUI begins

2006 2007 2008 2009 2010 2011 2012 2013 2014 20150.0

0.5

1.0

1.5

2.0

2.5

3.0

Whatever the ultimate architecture of future application processors, they will become more powerful and this will allow OEMs to create devices with new capabilities, or entirely new devices or platforms, which in turn are critical to triggering the next big success.

Immersive interfacesEnabling more immersive and natural user interfaces and experiences is one of the opportunities for creating an environment in which the next explosion can occur. Features such as 3D, gesture control and devices that process and act upon real world stimuli will be among the next evolutionary steps on this path, and will demand faster processors.

A new use case just beginning to take advantage of greater processor power is the ability to stream content wirelessly from a smartphone or tablet to a TV set or computer screen. Apple has launched AirPlay technology for content streaming among its products but not outside its ecosystem. For everyone else, OEMs are beginning to certify products on the Miracast standard, with some expected in stores for the holiday season. AirPlay and Miracast are based on newer versions of Wi-Fi (to be explored in a future article). In this content streaming use case, the application processor in the handheld is the pump that pushes the bits to the TV screen.

As this content model advances, more powerful processors will play another crucial role. Most of us quickly learn to operate our TV remote controls based on

the tactile feeling of the buttons while our eyes continue to watch the screen. As a remote control, the smartphone has virtual buttons, which users have to look at. Within our five-year forecast period, we expect a solution: new technology (to be explored in a future article) that senses where your fingers are hovering over the screen and projects an icon on the TV screen for the button you are above—channel changer or volume control, for instance. The sensing technology and software that will make this possible will require the level of processing power we are forecasting in future processors.

We anticipate the 1.66GHz threshold, which application processors will soon reach in large numbers, to be the launching point for the next cycle of innovation. [See Figure 2] It is important to keep in mind that not all of the next evolutionary steps will occur at the same time, but that this processor threshold is the starting point for some of these to begin to become viable in technology and cost.

In our forecast, the GHz/Core/$ metric enables devices to exceed this 1.66GHz threshold in large numbers in 2013-2014, and to approach the 3GHz threshold in 2015-2016. There is already at least one 2GHz smartphone that will be sold in markets outside North America before the end of 2012. Higher performing application processors at ever-declining prices will drive higher levels of smart handheld device penetration compared to feature phones, producing a larger total market.


© 2013 PwC. All rights reserved. PwC refers to the PwC network and/or one or more of its member firms, each of which is a separate legal entity. Please see http://www.pwc.com/structure for further details. AT-13-0059

Raman Chitkara Global Technology Industry Leader [email protected]

Pierre-Alain Sur Global Communications Industry Leader [email protected]

Chris Richard PwC Management Consultant Lead, Semiconductor Practice [email protected]

Let’s talk


As noted above, gigahertz increases in the smart phone are challenged by power limitations and heat dissipation problems that must be solved. While a 3GHz mobile processor is possible (at least one prototype exists), how the industry will surmount these physical hurdles is not clear.

“Even desktop processors have gotten out of the gigahertz race to some degree, delivering more cores and relying on parallelism to crank out more power rather than ever-faster clock speeds,” says Chris Richard, PwC Management

Consultant Lead, Semiconductor Practice. “Consumers will use all the gigahertz vendors can throw at them, but beyond 2015 it remains to be seen what scientists and engineers will be available to deliver.”

Chris Richard

PwC Management Consultant Lead, Semiconductor Practice



About PwC


1 Mobile Technologies Index Infrastructurespeed:Watchcapitalinvestmentin4Gforthenextinflection

www.pwc.com

Mobile Technologies Index Infrastructure speed:Watch capital investment in 4G for the next inflection

PwC predicts that infrastructure speed will be the fastest improving component of the seven enabling technologies in the PwC Mobile Technology Index through 2015.

By 2015, we also expect three factors associated with the transition to 4G technology—share of infrastructure investment, share of devices and share of subscribers—to reach levels that could trigger a robust period of 4G innovation. We base this expectation on the pattern we saw in the same three factors in the 2G-to-3G transition. As this pattern repeats with 4G it creates the potential for a surge of 4G innovation starting no later than 2015 [see Figure 1]. We expect this 4G innovation to include new business models based on capacity improvements, and new use cases based on better video streaming and other technologies. We explore all this later in the article.

By Pierre-Alain Sur, Global Communications Industry Leader

2006 2007 2008

3G innovation surge

4G innovatio

n surg

e2005

2013 2014 20152012

53%of 10%

of20%of

capital expenditures device penetration subscribers on 3G

49%of 10%

of23%of

capital expenditurespenetration

device penetration subscribers penetration

Figure 1: The shift from 3G to 4G will launch another innovation explosion

First, our forecast, and how we derived it.

The technology innovations that establish the speed at which data can travel to and from a mobile device happen in two places: the infrastructure speed capability outside the device, and the connectivity speed from the modem capability inside the device.[See Device connectivity article at www.pwc.com/mobileinnovations.] The Index includesbothfactorsasseparate,equally weighted components.

For myriad reasons, the actual speed for the user cannot be precisely predicted. At any point in time, the wireless infrastructure is a mix of technology generations. Carriers, for


good business reasons, don’t normally make available to end users the maximum speed the underlying technology is capable of supporting. For the user, the mobile experience is only as fast as the slowest point on the network at that moment.

But it is not the purpose of the PwC Mobile Innovations Forecast, of which the Index is just one part, to predict with precision the real speed, or even the average, that a user might experience in the future. The purpose is to understand and forecast mobile innovation. That is, to use infrastructure speed as we calculate it and all the other metrics in the Index to suggest direction and magnitude, both individually and collectively, and, in turn, to use these as vectors to help us identify patternsthatsuggestnewinflectionpointsformobile innovation.

For infrastructure speed, PwC forecasts a global CAGR through 2015 of 54 percent [see Figure 2], as measured in average Megabits per second (Mbps). This makes it the fastest improving component, just ahead of processor speed—53 percent CAGR in GigaHertz per dollar.

The forecast CAGR for infrastructure speed is less than the rate of improvement from 2007 to 2011 when it was a staggering 77 percent. The CAGR is slowing down mainly because 2007 era

speeds started off so slow; speeds in 2011 and beyond are in a different part of the S curve.

Estimating improvements in average infrastructure speed over time is a complex formulation, involving several judgment calls. First, it is important to recognise that the metric we use is really a measure of infrastructure speed capability, meaning the maximum speed at which a single device could communicate with the wireless infrastructure under optimal conditions—a bit likeautomotivefuelefficiencyratings.

In practice, the speed at which bits are streamed to individual devices by wireless infrastructure is determined by several variables. Among these are the limits placed on the network by the operators themselves for technical, service quality or business reasons. Other variables include geography, the generation of technology used by the cell tower, the number of other devices sharing the total capacity of the tower, the type of data the devices are accessing (text or video, for example) and the signal’s generation, strength and exposure to interference.

The goal of tracking the relative changes in speed from one year to the next makes this task a bit easier, with the caveat that average infrastructure speed refers to speed where wireless services are offered to begin with.

Figure 2: Infrastructure speed, compound average growth rate (CAGR)

2011–2015 CAGR = 54%

2007–2011 CAGR = 77%

0

100

200

300

400

500

600%

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Infrastructure Mbps

Per

cent

age

of 2

011

Ind

ex b

ase



With all of these factors in mind, we sought to create logic that resulted in an average network infrastructure speed capability across generations of technology and over time. Here is how we calculated the metric:

a) We sum the cumulative capital spend in any given year on each generation of wireless infrastructure [see Figure 3],

b) factor in “end of life” for infrastructure by discounting past investment by subtracting past years’ investments in 33 percent

increments beginning three years after their deployment,

c) and then create the weighted average network speed where the weights are the percentage of discounted, cumulative spend and the speeds are global industry values forspecifiedgenerations(3G,4G,etc.)[see Figure4].

To reiterate the earlier caveat, the speed data in Figure 4 are the anticipated optimal capabilities, not the actual speeds any user would experience.

0

10

20

30

40

50

60

70

80

90

100

2015$49.8

2014$48.1

2013$46.3

2012$44.9

2011$41.7

2010$39.1

2009$39.3

2008$41.5

2007$39.5

2006$38.4

2005$36.2

2004$33.8

2003$26.8

2002$27.5

Figure 3: Capex infrastructure market share chart

4G

3G

2G


Total carriers' annual spending (bn)

Figure 4: Infrastructure speed capability1 in megabits per second global average, 2006-2015


1. The maximum speed at which a single device could communicate with the wireless infrastructure assuming the service provider is not limiting these speeds for technical, service quality or business reasons.

Figure 4: Infrastructure speed capability in megabits per second global average, 2006-2015

0%

20%

40%

60%

80%

100%

120%

0

30

60

90

120

% c

hang

e

Mb

ps

Not normally seen in practice—theoretically highest speed infrastructure can offer.

2006 2007 201020092008 2011 2012 201520142013

Speed Percentage change


0

10

20

30

40

50

60

70

80

90

100

20157.3

20146.9

20136.5

20126.2

20115.8

20105.3

20094.6

20084.0

20073.3

20062.8

20052.2

20041.8

20031.3

20021.1

Figure 5: Subscriber market share

4G

3G/3.5G

2G/2.5G

1G


Total subscribers: (bn)

We now turn to a pattern that suggests the timing of a new market opportunity for mobile innovation. We considered both the infrastructure and the device. Understanding how these impacted 3G helped us to pinpoint a comparable 4G innovation opportunity.

In the device connectivity speed article we offered this rule of thumb for mobile innovation: when new capabilities reach a device penetration level of 20 percent, game-changing services can be launched and market disruption can ensue. The 3G modem technology reached device penetration of 20 percent in 2007 [see Figure 6].

That same year—2007—network operators’ investment in 3G infrastructure crossed the threshold of 50 percent of total capex (53 percent to be exact). Subscriber rates are a lagging indicator of innovation and not included in our Index, but in looking for patterns it is useful to note that 3G subscriptions reached 10 percent of the total in 2008 [see Figure 5].

Device penetration by 3G of 20 percent and the 50 percent threshold in 3G infrastructure capex spending, both reached in 2007, coincided with a surge in 3G application development in 2007 and 2008. Given this pattern, when might vendors expect a similar burst of 4G-based innovation comparable to the 3G surge? We expect it to start by 2015, if not sooner. Here is our reasoning.

Figure 6: Predicting the innovation surge

2006 2007 2008

3G innovation surge

4G innovatio

n surg

e

2005

2013 2014 20152012

53%of 10%

of20%of


49%of 10%

of23%of



2G to 3G began when penetration reached:

3G to 4G begins when penetration reaches:

2007 2007 2008

2013 2015 2015


For 3G, capex spending hit 50 percent and device penetration hit 20 percent in the same year—2007. Our forecast predicts 4G capex spending reaches close to 50 percent in 2013 (49 percent to be exact), but 4G device penetration lags, not reaching 23 percent until 2015. The 4G subscription rate reaches 10 percent in 2015, comparable to 3G in 2008.

Based on the 3G pattern, we expect a surge of 4G-based innovation that would start no later than 2015. If infrastructure capex or device penetration were to accelerate faster than forecast, the 4G-based innovation surge could start as early as 2014.

Whether it starts in 2015 or 2014, this surge of 4G-based innovation could trigger another shift in the wireless value chain’s major players, although such a shift would depend on some vendors recognising the same potential tipping point that we see, and seizing the opportunity ahead of others by creating innovations that 4G networks bring to life.

What kind of innovation can we expect? We anticipate that 4G innovation could spawn new use cases, involving more and better streaming video, including more satisfactory viewingofcommercialfilmandTVcontentfrom the cloud and multiplayer mobile gaming with minimal latency. Other use cases are likely to come in mobile video conferencing and voice-over-Internet services that rival or exceed the quality of traditional wire-line offerings; new device form factors better attuned to augmented reality; and other applications involving the movement of large amounts of information.

We also expect new vertical industry use cases. For example, when paired with improved image sensing,innovativenewsensorsandartificialintelligence, 4G could support new use cases such as remote medical diagnosis and repair

effortsbyfieldservicerepresentativesandbringback house calls by the family physician—in virtual form.

Some of these applications and use cases are possible even with 3G technology, but 4G will certainly accelerate their adoption by making them more widely available and by improving the user experience through somewhat faster downloads and lower latency. But whatever the new use cases are, we also expect the period of 4G innovation beginning in 2015 to differ somewhat from the 3G transition.

“To some extent, 4G may not impact mobile innovation the way 3G did,” observes Dan Hays, PwC US Wireless Advisory Leader. “We may be more likely to see second order effects from 4G rather than new things enabled by the technology itself. The use case and application innovation we saw with 3G may be less likely to recur than is business model innovation. I believe 4G will enable operators to deliver a more consistent experience, more ubiquitously, at a lower cost and allow them to make money and stay in business.”

We expect operators to take advantage of 4G capacity and speed to achieve service level expectations but not necessarily go beyond them. The burst of 4G innovation will also offer operators the opportunity to reduce the costs of network operations through improvements in workforce productivity and other innovation.

It might be that the network operators and their businessmodelscouldbenefitmostdirectlyfromthe4Ginflectionpointweidentify.Andyet, there is also another way to think about the 4Gtransition,basedonwhathappenedwith 3G.

2006 2007 2008

3G innovation surge

4G innovatio

n surg

e

2005

2013 2014 20152012

53%of 10%

of20%of


49%of 10%

of23%of



Figure 7: The 3G innovation zone occurred in 2007 and 2008

Dan Hays

US Wireless Advisory Leader

The initial transition to 3G started several years before the highly disruptive introduction of devices such as the iPhone, Android phones and tablets that took advantage of the newer technology. There were early great expectations, of course—new 3G-only carriers started in various places in the world with little initial business success because the user experience was not that different from 2G. This was before devices had the component power of a true mobile computer (the very components we track in the Index), and before the existence of application ecosystems that offeredcompellingnewuse cases.

Instead, the early 3G value proposition mostly appeared to be lower operating costs for carriers rather than disruptive new capabilities for users. The predictions today that 4G will once again reduce operating costs for carriers has good reasoning behind it, especially because the software in 4G infrastructure is automating more and more management functions.

But predicting that 4G will mainly reduce carrier operating costs would appear to deny the kind of impact from 4G that 3G eventually had in creating the setting for highly disruptive innovations. Our future research on 4G will explore this matter in greater detail.


© 2013 PwC. All rights reserved. PwC refers to the PwC network and/or one or more of its member firms, each of which is a separate legal entity. Please see http://www.pwc.com/structure for further details. BS-12-0541-C.0712.DvL

Pierre-Alain Sur Global Communications Industry leader [email protected]

Dan Hays US Wireless Advisory Leader [email protected]

Let’s talk



The Technology Institute is PwC’s global research network that studies the business of technology and the technology of business with the purpose of creating thought leadership that offers both fact-based analysis and experience-based perspectives. Technology Institute insights and viewpoints originate from active collaboration betweenourprofessionalsacrosstheglobeandtheirfirst-handexperiencesworkingin and with the technology industry. For more information please contact Raman Chitkara, Global Technology Industry Leader.

About PwC

PwCfirmshelporganisationsandindividualscreatethevaluethey’re looking for. We’re anetworkoffirmsin158countrieswithcloseto169,000peoplewhoarecommittedtodelivering quality in assurance, tax and advisory services. Tell us what matters to you andfindoutmorebyvisitingusathttp://www.pwc.com/.

1 Mobile Technologies Index Device Connectivity Speed: One half of an equation


Mobile Technologies Index Device connectivity speed: One half of an equation

From the user’s perspective, the mobile experience starts with the speed at which the device receives data and applications. That speed is, of course, the combined result of the speed capability of the modem technology inside the device, which is fixed, and the speed capability of the infrastructure, which can vary.

Thus, wireless speed is a complicated component to measure. So complicated, in fact, that we break it into two components, each with its own metric, in the PwC Mobile Technologies Index:

• Device connectivity speed in Megabits per second per dollar (Mbps/$)

• Infrastructure speed in average Megabits per second (Mbps)

In this article we offer our forecast for device connectivity speed, explain the metric and how we calculate it, and explore some implications for mobile innovation. In the next article we post [see, “Coming soon” at www.pwc.com/mobileinnovations], we will offer our forecast for infrastructure speed, and identify a pattern we see that involves both wireless speed components and that signals a future innovation inflection point.

PwC forecasts a compound annual growth rate (CAGR) of 37 percent for average aggregated device connectivity speed as measured in Mbps/$ through 2015. [see Figure 1] Put another way, average aggregated device connectivity speed will be four times greater in 2015 than in 2011.

Figure 1: Device speed capability CAGR, 2011–2015

0

.5

1.0

1.5

’10 ’11 ’12 ’13 ’14 ’15’09’08’07

75% CAGR (2017–2011)Moving from 2G to 3G

37% CAGR (2011–2015)Moving from 3G to 4G


Mbps/$

By Pierre-Alain Sur, Global Communications Industry Leader


The device connectivity speed metric is actually an aggregation of metrics for all wireless generations in use, plus the improvements we anticipate for each generation during their period in use.

Device connectivity speed is defined as the theoretical maximum speed at which a mobile device can operate using a particular air interface technology, also known as a generation of radio transmission. The theoretical speed of each air interface is fixed, but the average speed realised within each generation of technology can improve through the optimisation of handsets, base stations and air interface protocols.

In addition to these variables, average aggregated speed in each generation will improve as OEMs shift production mix to faster protocols that exist within the generation. Shifts within generations will deliver the following improvements in Mbps/$ from 2011 to 2015:

• 2G speeds will improve at 3 percent CAGR as the mix moves from GSM to Edge.

• 3G speeds will improve at 6 percent CAGR as mix moves from EVDO and WCDMA to HSPA+.

• 4G speeds will improve at 8 percent CAGR as the mix moves from early WiMAX to mature LTE.

These incremental gains may appear to be modest, but there is another dynamic at work: the mobile device production mix shifts as the industry moves from 2G to 3G to 4G. So we take

the average speeds and weight them based on mobile handset production for each generation, and wind up with the Mbps/$ of “total devices produced” each year improving at a 55 percent CAGR, 2011 through 2015.

However, this 55 percent CAGR omits the cost of the main components in a handset that enable communication over the multiple air interfaces that end devices must support. After factoring in these costs, the actual Mbps/$ in the next four years will increase by 37 percent—which is the device connectivity speed CAGR we use in our Index.

This is just half the rate of improvement we saw in the period 2007-2011, when the CAGR was a staggering 75 percent. The slower increase in Mbps/$ is primarily due to the baseband chipset costs being significantly greater for the move from 3G to 4G than they were in earlier transitions.

Nonetheless, we anticipate that 37 percent CAGR is enough improvement to enable continuing mobile innovation at a rapid pace. A review of recent history explains why we are confident in saying this.

When the original 2G iPhone was launched in 2007, Apple proved that consumers would accept iTunes on handsets and the Apps Store concept. However, 2G connectivity was slow enough to risk failure for the original iPhone if the faster 3G version had not been launched one year later. By the time the 3G iPhone was introduced, the mobile handset supply chain was already producing more than 200 million 3G phones per year. Apple didn’t have to create the 3G technology, it just had to put it to use.

By the time Apple needed a faster connection to support its vision, the electronics industry was already devoting 20 percent of production to 3G handsets. Consumers were buying the faster handsets before many applications existed that needed the speed, and in many cases before infrastructure speed had fully transitioned. In contrast to classic “pent-up demand,” consumers were priming the pump for demand by pre-purchasing capabilities ahead of actual use cases, with the expectation that when 3G applications and 3G networks were available, their handsets were ready.

37%Annual increase in mobile device connectivity speed


We anticipate this trend will continue, and we use this as an example of the following rule of thumb for timing the launch of a disruptive mobile venture:

When new capabilities reach a penetration level of 20 percent, game-changing services can be launched, and market disruption can ensue. At the 20 percent level, the market has begun transitioning from a relatively few early adopters to a mass market, and the entire ecosystem, including new entrepreneurs, are developing and positioning for game-changing solutions.

In 2007 and 2008, 20 to 25 percent of device production was dedicated to 3G [see Figure 2]; this period coincided with the initial surge of 3G applications development. From the standpoint of infrastructure capital expenditure, 3G coverage had reached at least 53 percent of what carriers were spending on 2G during that time. (We will say more about this in the upcoming infrastructure article.)

All the 3G-related factors—device production, applications development and infrastructure investment—set the stage for the success of Apple’s iPhone, Google’s Android, apps stores and other key mobile device phenomena that have contributed to the mobile ecosystem as we know it in 2012. 4G will offer even faster speed and less latency, which makes the speed more useful.

Together, improved device connectivity speed and improved infrastructure speed will deliver another wave of innovation and disruption (to be further explored in the next article). The move from 3G to 4G will enable new business models for carriers, and new use models for the mobile device, including more and better streaming video, mobile video conferencing, voice-over-Internet services and other applications involving the movement of large amounts of information, including the growing mass of data collected by the mobile device itself and transferred wirelessly to the cloud for analysis, and back again for action by the user.

Figure 2: 4G poised to drive the industry in 2015

Device Production

2007 2008 2009 2010 2011 2012 2013 2014 2015CAGR

(2011–15)

2G 80% 75% 71% 66% 58% 51% 44% 37% 31% -14%

3G 20% 25% 29% 33% 41% 45% 47% 49% 45% 3%

4G 0% 0% 0% 0% 1% 4% 9% 14% 23% 121%



© 2013 PwC. All rights reserved. PwC refers to the PwC network and/or one or more of its member firms, each of which is a separate legal entity. Please see http://www.pwc.com/structure for further details. BS-12-0541-B.0612.DvL



About PwC




Let’s talk




The PwC Mobile Innovations Forecast Making sense of the rapid change in mobile innovation

With a quick boot, instant response to touch and speedy downloads from the cloud-based App Store, the application processor was one of the iPhone’s many competitive advantages when it debuted in 2007. One factor contributing to the powerful processing capability was Apple’s decision to use NAND flash memory instead of NOR. At the time Apple made that decision, NOR was the standard flash technology used in mobile phones and NAND was an emerging technology—faster and denser than NOR, but more expensive. For those performance reasons, Apple chose NAND, taking a calculated risk that the price-performance would improve, due in part to the demand that Apple itself would create with a successful launch of the iPhone.

Within two years, NAND became the standard, not only for system boot up but also for storage.

“Looking back, one overlooked but key enabling technology for the iPhone was moving from NOR flash to NAND flash,” says Steven Mather, Senior Principal Analyst at IHS, a global information and analytics provider. “You needed a powerful operating system and processor, but you couldn’t do that until you had the memory, and you couldn’t do the memory until you had the NAND flash. NOR wasn’t capable of supporting a higher level apps processor like the one in the iPhone.”

NAND was just one catalyst for the iPhone, which became one of the most successful high-tech product launches of all time.

“The team at Apple recognised changes in technology, usage and materials,” says Dan Hays, PwC US Wireless Advisory Leader. “Designing a phone that was that thin with that battery life wasn’t as much about beating the others to market. It was about recognising fundamental changes in materials, technology and things that you could do with industrial design and user interface.”

Dan Hays

PwC US Wireless Advisory Leader

The point is this: Apple’s decisions about NAND and other components illustrate how understanding the evolutionary curve of technological innovation, even of commodities like flash memory, can lead to a disruptive product that transforms an entire ecosystem. Where will the disruptions in mobile innovation arise over the next five years? How will they change consumer and employee behaviour? What business opportunities will result? What can companies do to take advantage of these disruptions? How do they fit into broader market trends now driving the technology sector?


2 The PwC Mobile Innovations Forecast Making sense of the rapid change in mobile innovation

Introducing the PwC Mobile Innovations Forecast

Answering these kinds of questions requires not just a keen understanding of the evolutionary curve of the enabling technologies, but a broader framework for analysing mobile innovation quantitatively and qualitatively. So with the goal of providing business leaders early warnings about coming disruptions and actionable intelligence about new opportunities, PwC introduces its Mobile Innovations Forecast (MIF), a four-part framework for analysing and understanding mobile innovation. The four parts are:

• Enabling technologies;

• New technological capabilities;

• New use cases and

• New business models.

The four parts will be explored in periodic articles on this Web site in the months ahead. We expect that examining, analysing and forecasting mobile innovation along these lines will shed light on the interdependencies that are otherwise cloaked by the unorganised daily stream of innovation announcements from the mobile ecosystem.

The first category—enabling technologies—is the focus of PwC’s Mobile Technologies Index, a new quantitative method developed to analyse the rate of improvement in key technologies that are fundamental to mobile innovation, and to help forecast new use cases and business models. (see sidebar, “Creating the Mobile Technologies Index, page 8) These enabling technologies have led to the rapid improvement of mobile device capabilities, dramatically changing our personal and work lives, and the way several billion people interact with each other.

Over the next five years, the smartphone will continue to acquire capabilities that will make it more and more like a full-fledged personal computer, but in the same period mobile innovation will continue to extend beyond smartphones and tablets. Mobile innovation in health care, automotive, home entertainment, manufacturing and other diverse sectors is likely to be just as robust.

“We will track things like the number of technologies, how fast are they moving and how that enables innovations. If you get so much more processing speed and you get so many more features then it allows you to do something different,” says Rodger Howell, PwC Principal for Mobile Computing. “We will monitor progress along this vector, and we will also note when other vectors are showing up, new threads, new business models, new use cases—here’s a concept that somebody’s trying in a market not proven yet.”

Rodger Howell

PwC Principal for Mobile Computing

This article, the first of many, introduces the PwC Mobile Technologies Index; then explains where the Index fits in the four-part framework of the Mobile Innovations Forecast, and concludes with a look at how mobile computing contributes to the market and industry forces that are driving the broader technology sector. PwC identifies mobile computing as one of four market forces that are individually and collectively redefining customer demand, expectations and business opportunity. The others are cloud computing, social technology and the emergence of intelligent devices (the digitisation of inanimate objects). Together, these mega-trends are leading us toward an era of ubiquitous computing, which relies especially on wireless networks.

New capabilities

The Mobile Innovations Forecast includes technologies that are not currently significant enablers of innovation, but could become important within our five-year period. In our four-part framework we categorise these as new capabilities and they will be the focus of a future report. [See Coming soon at www.pwc.com/mobileinnovations]

“As important as the enabling technologies in our Index are, some of the more interesting use cases and disruptive mobile innovation are likely to be driven by the emerging technologies and new capabilities of existing technologies, which we include in this group,” says Daniel Eckert, a PwC Director for Mobile Computing.

Daniel Eckert


New capabilities include technologies that could change how users interact with the devices and how the devices interact with the environment. Based primarily on qualitative research to date, we can suggest several examples we might include: near-field communications (NFC), high-definition audio (HDA), 3D computing, future generations of voice recognition, artificial intelligence, advanced video compression, gesture sensing and olfactory sensing (artificial nose).

Based on our historical understanding of technology adoption, new capabilities for the purpose of our forecast framework are those that have not yet met the threshold of 20 percent penetration of mobile devices but are likely to do so within our five-year timeframe.

One such example is NFC, which allows for secure, simplified transactions, data exchange and wireless connections between two devices near each other—known as proximity detection. NFC’s

— Continued on next page


Figure 1: PwC Mobile Technologies Index

41% CAGR (2011–2015)


55% CAGR (2007–2011)

0

100

200

300

400%

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

Mob

ile c

apab

ilitie

s as

a p

erce

ntag

e of

201

1 In

dex

bas

e

The Index: 41% CAGR through 2015

Our examination of mobile innovation begins with the key enabling components, introduced here and then analysed in more depth, component by component, in separate articles in the weeks ahead. For this purpose, we have created the PwC Mobile Technologies Index, a broad composite of seven enabling components that underlie the power of the mobile device to sense, analyse, store and connect information. Since the first brick-like mobile phones began to appear in the chauffeured limousines of business executives and movie stars in the 1970s, disruptive breakthroughs in mobile have resulted due, in part, to the continuous progress of these components at predictable price points. Thus, our forecast begins with them.

In this first release of the Mobile Technologies Index, PwC forecasts a combined compound annual growth rate (CAGR) of the Index between 2011 and 2015 of 41 percent. [see Figure 1, “PwC Mobile Technologies Index”] As Figure 1 shows, this is less than the 55% CAGR of the Index between 2006 and 2011, but still represents large enough improvements in the underlying components to anticipate many new mobile value propositions.

Here is the forecast through 2015 for each of the seven enabling technologies in the Index:

• Infrastructure speed: In average Megabits per second (Mbps), will improve 54 percent CAGR.

• Device connectivity speed: In Megabits per second per dollar (Mbps/$), will improve 37 percent CAGR.

• Processor speed: In GigaHertz per dollar (GHz/$), will improve 53 percent CAGR.

• Memory: In Gigabits per dollar (Gb/$), will improve 48 percent CAGR.

• Storage: In GigaBytes per dollar (GB/$), will improve 35 percent CAGR.

• Image sensor: In Megapixels per dollar (MP/$), will improve 20 percent CAGR.

• Display: In performance per dollar per square inch (P/$/in2), will improve 16 percent CAGR. (Performance is a weighted aggregation of resolu-tion, brightness, power efficiency and other factors.)

device penetration reached an estimated 7 percent by the end of 2011, but we expect it to hit the 20 percent threshold within a couple of years. NFC has the potential to drive use cases and business models around device-based electronic payment systems as an alternative to credit cards (already happening in parts of rural China), to being a proxy for tangible keys, money, tickets, travel cards and identity documents.

Based on our future research, there could be several micro-electro mechanical systems (MEMS) that we would include as new capabilities. At present, the most widely adopted MEMS device is the accelerometer, which rotates the smartphone screen from vertical to horizontal. Its device penetration rate increases from an estimated 45 percent by the end of 2011 to an estimated 69 percent by 2015. Compasses, gyroscopes and pressure sensors are three other MEMS devices we are tracking with data. Compasses reached an estimated 20 percent by the end of 2011, and will hit an estimated 44 percent in 2015, but the other two do not reach 20 percent by then.

New versions of Wi-Fi and Bluetooth currently under development could re-energise these technologies and make them “new capabilities” once again. A new version of Wi-Fi will offer ultra-high bandwidth for line-of-sight or in-room applications. A new version of Bluetooth is anticipated for ultra-low power applications. When these new versions appear, we would likely cover them as new capabilities.

We also include power and batteries under new capabilities. A smartphone that only needed recharging once a week would be a game changer. Batteries have improved only gradually for the past decade and are expected to continue their relatively slow advance. New capabilities that would accelerate improvements in mobile device batteries would create new opportunities, so our forecast will be sensitive to major disruptions emerging from innovation in power management and battery life.


You might notice that the Index does not include battery or power management. This is not an oversight, but an informed decision. Our research indicates that battery performance improvements have been limited in nature and are not currently forecasted to be anything close to what we have seen for the components we chose to include in the Index, such as processor speed or storage. We will keep a close watch on battery technology; if there is significant change in the offing we anticipate that it will be covered in future articles describing new technological capabilities. (see sidebar, “New capabilities”)

The Index from an historical perspective

The price-performance improvements of these enabling technologies are largely responsible for the current trend in which smartphones are gaining market share over feature phones. But all mobile devices become cheaper, faster, more

capable sooner, not later, and are able to deliver more, better and more diverse services and digital content of all kinds.

“The seven components together, on average, will improve 41 percent a year, each year, for the next few years. So the kind of dramatic change from phones we had in 2006 to the iPhone in 2007 is going to continue,” IHS’s Mather says.

By calibrating each component at 100, with 2006 as the base year, the spider diagram illustrates how the seven components are progressing relative to each other. [see Figure 2]

In the weeks ahead, we will examine these technologies in a series of articles, including one that looks at the operating system (OS), which is not included in the Index, but is a key enabler of mobile innovation. [See Coming soon at www.pwc.com/mobileinnovations] The first of these articles, looking at device connectivity speed, is available here.

Processor speed

Memory

Device connectivity speed

Infrastructure speed

Display

Storage

Imaging sensor

2006—2015

2006—2011

100 100200 200300 300400 400500 500600% 600%

Figure 2: PwC Mobile Technologies Index—Relative progress of components

Note: Infrastructure speed, Device connectivity speed, Storage, Memory and Processor speed are core technologies.Each accounts for 16% of the Index. Imaging sensor and Display each account for 10% of the Index.See “Creating the Mobile Technologies Index” on page 8 for more information.

Source: IHS iSuppli Databases

New use cases

The technologies that comprise the Mobile Innovations Forecast and new capabilities are important, but some of the most interesting questions about mobile innovation centre on evolving use cases.

Tracking rates of change in various technologies will help us make modest predictions about use cases already under development. For example, within the five years of our forecast period we anticipate the smartphone will add new and improved features that will give it the power and capabilities of today’s desktop and laptop computers.

“I’m carrying in my pocket right now a smartphone that has, on the size of my fingernail, 32 GB of storage. My laptop probably has eight times that,” says Dan Hays, PwC US Wireless Advisory Leader. “But it’s not farfetched to ask why shouldn’t all that storage just be on the phone that I carry around with me all the time. It’s secured and why shouldn’t I just dock it into a bigger screen when I want a tablet or dock it into a shell when I want a laptop or dock it into my TV at home when I want to watch movies.”

The extent to which the smartphone actually disrupts personal computers remains to be seen. But consider this: the Apple App Store now offers more than 650,000 apps,1 and the number of Android apps has topped 450,000.2, 3 This continuing proliferation of apps suggests any number of future mobile use cases that will extend the power and scope of mobile devices.

1 http://www.bgr.com/2012/06/11/apple-kicks-off-wwdc-2012/

2 http://androidcommunity.com/android-growth-continues-450000-apps-850000-activations-daily-20120227/

3 http://www.engadget.com/2012/02/27/google-450-000-android-apps-now-available-to-300-million-device/



The four-part framework of the Mobile Innovations Forecast

As noted, the Mobile Technologies Index is just the starting point for our ongoing forecasting efforts in mobile innovation. Wireless devices and their supporting services will continue to run applications faster, store more data, create better pictures, and display information in brighter and more compelling images, driven by the seven components of the Mobile Technologies Index. “The seven components, individually or collectively, are not likely to cause the next great disruption in the next five years without some creative thinking about how to use them. Specifically, new capabilities, cre-ative use cases and imaginative business models—or some combination—are more likely to produce a game-changing mobile innovation,” says Kayvan Shahabi, PwC US Technology Advisory Leader. We will more closely examine these three areas in future reports on this Web site.

Kayvan Shahabi

PwC US Technology Advisory Leader

“People are focused on the things they know or are easy to understand, but those might be the wrong things,” PwC’s Hays cautions. “There are two pieces to this story. There’s how things are evolving and tactically moving up the curve and then there’s the big gotchas.”

To anticipate the gotchas and track their progress from idea to commercial reality, we constructed the broader framework of the Mobile Innovations Forecast. The Mobile Technologies Index tracks innova-tion that, in aggregate, is analogous to a rising tide that lifts heavier boats still resting on a muddy bottom. Predictably, future component performance levels at acceptable price points will support even “heavier” applications that are not fea-sible today. But other types of innovation are harder to predict and the consumer, and even some established vendors, can find it hard to anticipate them.

“There’s a segment of just new ideas that are going to come in. We don’t know what they are, but people are working on con-cepts that may or may not take off,” PwC’s Howell says.

A domain of technology with this amount of innovation poses a challenge for apply-ing a useful framework on its evolution: which attributes should be included? Add the complexities of personal, enterprise, national, service provider and global perspectives, and the problem becomes nearly unsolvable. The framework must be flexible enough to account for continu-ous introduction of the new. And it must separate innovations into meaningful cat-egories while retaining an explicit expec-tation of emergent interdependencies.

The historical context must also be considered. There was a time when most functions now standard in smartphones were too expensive or unusable due to the immaturity of the underlying technolo-gies. Such functions included integrated imaging, general purpose operating sys-tems, downloading applications, location awareness, motion sensing and others. How long after the technical capability existed did it take for these to become mainstream features, and what happened that allowed them and the use models they support to become standard?

From this perspective, more interest-ing questions about the future of mobile devices can be considered. What features and functions are coming, but are not commercially possible today due to cur-rent technology limitations, cost, wireless network speeds, business model imperfec-tions or other barriers? What constraints prevent their appearance? When will these constraints dissipate enough to allow the capability to flourish? Can we predict when mobile devices will incor-porate potentially disruptive capabilities they don’t have today, based on future engineering advances?

The smartphone is already causing disruption in several categories of electronics—alarm clocks, digital cameras, gaming devices, audio players and GPS devices, to name a few. Smartphone penetration in other use areas will also become predictable.

But if we only deal with obvious use cases entirely predictable from technology trends, then we will be limited to simply forecasting improvement in those uses over time. Our future research will lead us to discover and examine ideas still in the pure concept stage. Some of these seeds, planted over the next five years, are likely to come to fruition.

Such use cases will be found in numerous areas: health sensors and medical applications from on-person monitoring to alert systems; the electronic wallet; the automobile as an evolving mobile computing platform; wearable computing; highly evolved personal assistants. We can anticipate new classes of business use cases as well.

When we take a closer look at use cases in a future report, we will answer two key questions:

• What use cases are still on the shelf because of unsolved problems—technical or otherwise—and which of those problems have a chance of being solved in the next five years, thus unlocking the use case.

• What use cases not even on the shelf yet could move quickly ahead if some technology or business model problem is solved.

The answers will come from the use cases we identify through our in-depth interviews and qualitative analysis.


In this ongoing series of articles, PwC will periodically offer answers to those types of questions through the frame-work of the four categories of the Mobile Innovations Forecast:

• The seven enabling technologies of the Mobile Technologies Index, plus the OS, as explained above, are the subject of this current report and a series of articles to appear in the coming weeks.

• The new capabilities of various emerg-ing and existing technologies. [see sidebar, “New capabilities,” and the subject of a series of articles scheduled for later this year]

• New use cases that arise from perfor-mance improvements or entirely new mobile technologies including the extension of the mobile ecosystem into the cloud. [see sidebar, “New use cases,” and the subject of future articles]

• New business models built on all of the above, and that might increasingly rely on industry dynamics outside of the mobile industry itself. [see sidebar, “New business models,” and also the subject of future articles]

We do not intend to develop an encyclo-pedic description of all existing or poten-tial capabilities, emerging technologies, use cases and business models. In many cases we expect there will be an explicit hierarchy in the new—a new capabil-ity, relying on further improvements in enabling components, creates the poten-tial for new use cases that, in turn, create new business models.

The broader business impact

As noted, this introductory article and the individual component articles in the weeks ahead are just the beginning of a much longer project to track mobile innovation across a broad front. One hope is that, over time, the community of read-ers will offer their own ideas and insights into this process, and begin to understand how mobile innovation itself is only one key element in a broader evolving technology ecosystem.

The Mobile Innovations Forecast exists within PwC’s framework for understand-ing various dynamics driving the broader technology sector today, a framework that suggests ways technology compa-nies might navigate disruptions that are rich in opportunity. In this context, there has been little research that parses the mobile ecosystem as we are parsing it, and then analyses the ecosystem parts to forecast innovations likely to reach commercial adoption.

It is a truism that technology innovation never stops, and technology companies can never rest on prior success. In this regard, the disruptive forces in mobile computing are familiar—as friend and foe. But there is something different this time. PwC sees major market and industry forces co-mingling in ways that paint a forward-looking picture that is starkly, even radically, unlike the past. Incremental change this is not.

As noted earlier, mobile innovation is one of four market forces that, individually and collectively, are redefining customer demand, expectations and business opportunity for technology companies; the others are cloud computing, social technology and the emergence of intelligent devices. Individually, each is turning the rules of the broader technology sector upside down. [see Figure 3] For examples, consider the following:

• Enterprise computing device strategies focused solely on desktop PCs and lap-tops ignore at their peril the 24x7 real-ity of fully engaged knowledge workers and customers on mobile devices.

• The lack of agility in typical legacy data centres destroys enterprise value rela-tive to cloud options.

• Email and Web portal strategies are siloed and tone-deaf if they aren’t embracing many-to-many social technologies.

• Networked, digital intelligence will be emanating from billions of smart, networked “things” operating without direct human oversight and offer-ing data for purposes limited only by the imagination.

New business models

When, where and how will the next disruptions to mobile business models happen? Perhaps a social networking site becomes a virtual network operator? Or a healthcare provider resells wireless devices and services as part of a chronic disease management solution. Maybe a carrier launches vertical service and application packages to industry segments. We’re already seeing announcements by carriers and credit card companies that portend major business model shifts.

How will the cloud figure in mobile business models? Does the newly emerging wholesale wide area network “pipe” business make sense? Will more electronics retailers offer branded devices and service? How far below the $100 price point might the full-featured smartphone drop?

These are among the topics we will examine when we do additional research into business models in future reports. We anticipate considerable innovation over the next five years.

“You could argue that today there’s very little differentiation in the mobile communications industry,” says Dan Hays, PwC US Wireless Advisory Leader. “The phones are all starting to look alike. The services are all becoming fairly ubiquitous. It all feels very similar. That’s when there’s the potential for business model disruption.”

Operators, OEMs, retailers, automobile companies, healthcare providers and insurers, web-based entrepreneurs, the entertainment industry and any number of segments we haven’t considered yet, will seek new ways to monetise mobile innovation. New business models have the potential to disrupt as much as technologies do.



“Mobile is one of several disruptive changes affecting technology, communi-cations and media industries. The others being cloud computing, social media and the network of intelligent devices through the Internet,” says Vicki Huff, Technology Industry Partner based in Silicon Valley. “The individual impact of each could threaten established vendors and create new customer value propositions—some-thing mobile is already demonstrating in spades. But their combined impact is likely to be greater as mobile devices engage with smart objects without user intervention, incorporate personal data stored in the cloud and socialise com-merce. We expect the same level of disruption we have seen in the mobile ecosystem to play out in all corners of the technology industry, in some cases bring-ing former competitors together and in others turning friends into rivals.”

PwC views these four trends as delivering on the long-time promise of ubiquitous computing, a phenomenon predicted two decades ago by Xerox PARC. Today, Wikipedia captures it well: “Ubiquitous computing (ubicomp) is a post-desktop model of human-computer interaction in which information processing has been thoroughly integrated into everyday objects and activities.” When our custom-ers and employees live and work in this ubiquity they have disruptive expecta-tions that their work lives will be as

“thoroughly integrated” as their private lives—hence the “consumerisation of IT” that every enterprise is grappling with today.

Vicki Huff

PwC Technology Industry Partner

As if these market forces weren’t enough, the tech industry itself is experiencing major internal disruptions: maturity and convergence, globalisation, patents as an arms race and digital transformation of business ecosystems. Market forces are driving some of these disruptive develop-ments. Cloud totally redefines compute infrastructure. Everyone is in everyone else’s markets—hardware companies go into retail and services, while retail-ers introduce their own hardware and software companies become retailers and hardware vendors. And mobile comput-ing is establishing heated competition for the future of the end-user devices, a battle that reverberates throughout the value chain, from chips to applications. Inevitably the migration of major powers towards overlapping customers is result-ing in IP battles, even as globalisation is bringing new players onto the land-scape—largely made possible by digitally transformed product development teams,

Figure 3: The interplay of market and industry forces and how they are driving technology companies to consider new business models and more agile operations

Source: PwC

IntelligentDevices

CloudComputing

Social Technologies

PatentWars

GlobalFocus

Maturation/Convergence

DigitalTransformation

Consumerisationof IT

Ubiquitouscomputing

Pressure to bemore agile

New businessmodels

MobileComputing

Industry ForcesMarket Forces

— Continued on page 9

For example, when Apple launched the iPhone in 2007 it injected a new dynamic to the OEM and wireless carrier relationship, changing the standard business model. For the first time a carrier allowed an OEM to dictate design and functionality of the handset because it believed the innovations would jumpstart the adoption of its 3G service.

Concerned about being disrupted by others, wireless operators in particular are looking to evolve business models. “The operators are wrestling with how they can add value to avoid becoming just a dumb pipe,” says Jagdish Rebello, Director of Consumer and Communications Research at IHS. “In this, operators are competing against other operators, but they are also competing with content providers, OEMs, app developers and other nodes.”

The technologies in the Mobile Technologies Index and in new capabilities will suggest new use cases, which in turn can inspire new business models. And the whole cycle is likely to turn around on itself when new business models create demand for or accelerate the development of new technologies.

Business models are not part of the Index, which is why our future analysis of them will rely on interviews with leading industry visionaries and other research to identify and analyse where new business models are possible, are being tried or thought about and where current business models are most likely to be disrupted.


Creating the Mobile Technologies Index

The Mobile Technologies Index is a method PwC developed to analyse the rate of improvement in key enabling technologies that are fundamental to mobile innovation, and to help forecast new use cases and business models.

In the spring of 2011, PwC began to research whether there was a way to forecast mobile capabilities. In addition to seeking credible data about the technological drivers of mobile capabilities, we sought a shared vision that mobile capabilities could be forecast. After an extensive review and discussion with various market research firms, PwC partnered with IHS.

IHS, an Englewood, Colorado-based global information and analytics provider, has a comprehensive database of each sector of the high technology industry value chain, from analogue and digital circuit designers through the chip foundries, as well as the original equipment manufacturers and the telecom network operators.

More than 130 subsectors of the high tech industry are included in its databases of prices, production capacities, capital spending, component features, average selling prices and numerous other metrics. More than 140 analysts have compiled these historical data points going back 10 years, and have prepared forecasts extending through 2015. The IHS iSuppli products and services include information from teardowns of various phones and tablets over the years, along with a continuing series of surveys, vendor roadmaps, market share analyses and other sources.

PwC mobile computing and high tech industry practice leaders and IHS analysts reviewed the databases and identified several metrics that drove mobile capabilities in the past, and would be expected to drive the devices and services in the future. We eventually chose seven components for the Mobile Technologies Index: device connectivity speed, infrastructure speed, processor speed, memory, storage, display and image sensor. [See story, “Making sense of the rapid change in mobile innovation”]

Through a series of back testing and other calculations, we determined that the relative performance of these Mobile Technologies Index components has consistently followed a pattern of at least 30 percent compounded annual growth. The components, and therefore the Index’s performance, appear likely to continue in the same upward path for the next five years.

In addition to the analysis of the data, PwC practice leaders interviewed leading mobile industry visionaries to review the data and provide their perspectives on the future capabilities of smartphones, tablets, wearable computers and other portable devices.

We concluded that five technologies will continue to serve as the basic building blocks of mobile innovation: device connectivity speed, infrastructure speed, processor speed, memory and storage. Over the five-year forecast period, each of the five will continue to progress along a Moore’s Law type price-performance curve, which makes them a natural starting point for a numerical and predictive index.

Based on our analysis, we added two other technologies to the Index: display and image sensor. Both are progressing along a Moore’s Law-style price-performance curve, which we expect will continue for the five-year forecast period. Currently both are nearly as important to mobile innovation as the five core technologies.

The display, with its multi-touch capability, has been one of the most disruptive qualities of smartphones. Over the next five years we anticipate that touch sensitivities will improve, and the physical display will become thinner, tougher and less expensive per square inch. Resolution and power efficiency are also expected to improve. As for the image sensor, not only will the smartphone continue to disrupt the digital camera market, this technology will be integral to next generation social networking, which is expected to consume video the way today’s social networking consumes still photos.

We decided to weight these two at less than the value of the five core technologies. Each core building block technology accounts for 16 percent of the Mobile Technologies Index, and image sensor and display account for 10 percent each.

We examined several other technologies, some with meaningful metrics and some without. Among the more important ones were Wi-Fi, Bluetooth and operating systems (OS).

Wi-Fi and Bluetooth have already achieved wide adoption. By the end of 2011, Bluetooth was found in an estimated 81 percent of handsets and its estimated CAGR is only 3 percent in Mbps/$ through 2015. Sixty percent of handsets will have Wi-Fi capability by 2015, including virtually 100 percent of smartphones. Its CAGR is 18 percent Mbps/$ through 2015. Given these factors, we decided to exclude them from the Index.

We would like to have included the OS in the Index, but we have observed that the OS does not trend along a predictable evolutionary curve similar to the other components. Rather, it tends to lurch forward in disruptive ways with qualitative enhancements to the user experience. User experience can be measured through usability testing, but such studies of mobile devices are not widely conducted and tend to test for basic functions, making them a bit simplistic. The OS can also be benchmarked for standard criteria, but we are unaware of anyone issuing benchmark or usability data covering the past five years or forecasting the next five years for mobile operating systems.

Despite these barriers to including an OS metric of some sort in the Index, the OS is central to innovation, and so we will include a qualitative analysis of it in this first series of reports, which are focused on the enabling technologies.

© 2013 PwC. All rights reserved. PwC refers to the PwC network and/or one or more of its member firms, each of which is a separate legal entity. Please see http://www.pwc.com/structure for further details. BS-12-0541-A.0612.DvL


Tom Archer US Technology Industry Leader [email protected]

Vicki Huff Technology Industry Partner [email protected]

Kayvan Shahabi US Technology Advisory Leader [email protected]


Daniel Eckert Mobile Computing Director [email protected]

Dan Hays US Wireless Advisory Leader [email protected]

Rodger Howell Mobile Computing Principal [email protected]

Let’s talk


supply chains distribution channels and customer experience management.

Given this landscape, PwC’s Tom Archer, US Technology Industry Leader, says that all eyes should be on dynamic business model evolu-tion. “Vendors are looking at business models in a much more fluid way—across a spec-trum—anchored by products at one end and experiences at the other. Increasingly, vendors are positioning themselves somewhere in the middle: product with services; product with experience; services/experience with prod-uct; and service. This is easier said than done, of course, which means successful business model evolution requires a real investment in enterprise agility. Customers are already vot-ing with their dollars—with mobile computing leading the way.”

Tom Archer

US Technology Industry Leader

By examining mobile innovations in the framework outlined above, PwC hopes to cre-ate common understandings from which will emerge the launch points for disruptive innova-tions. In the weeks and months ahead, we will offer forecasts of the innovations positioned for near-term commercial success and around which others can build their own value proposi-tions. We have not reached the end of dramatic changes through mobile innovation. We are only at the beginning.



About PwC



Mobile Technologies Index Mobile operating system: Smartphones will just get smarter

The smartphone seems to acquire more cognitive capabilities with each major product release—regardless of the OEM. PwC expects this trend to continue.

Many of these cognitive improvements will be enabled by the technologies that comprise the PwC Mobile Technology Index—device connectivity speed, network speed, application processor, memory (DRAM), storage, image sensor and display. The mobile operating system (OS) is not included in the Index because there is no comprehensive metric for it, but we do anticipate continuous innovation, especially in the user interface (UI) layer and the core services layer (capabilities that make life easier for application developers) of the OS, to support the new cognitive capabilities of mobile devices.

Cognition is defined as ‘the mental action or process of acquiring knowledge and understanding through thought, experience, and the senses.’1 Propelled by sensor technology, more powerful processors, better connectivity, cloud data sources and increasingly sophisticated data analysis on and off the device, smartphones, tablets and other mobile devices are on track to acquire ever more of this capability. “The smartphone, in particular, appears destined to become a true digital assistant, capable of self-learning based on the user’s behaviours, then personalising both the active and passive ways the device can engage with the user, the environment and network,” says Daniel Eckert, PwC director for Mobile Computing.

Daniel Eckert


The role of the OS will only grow in importance as the orchestrator of all the components and services on the device and those that reside in the cloud. The OS is the enabler of the enablers, if you will.

Although there is no comprehensive metric for the OS, we see incremental improvement over time similar to DRAM, storage and the other mobile building blocks. Unlike the other components, it is difficult to define metrics for OS performance. Instead of more bits transferred per second, or more bytes stored per dollar, OS improvements come in qualitative enhancements: better security, multi-tasking capability, support for more media protocols, etc.

One reasonable approach to forecasting change in the OS is to examine the history of these incremental improvements and look for patterns likely to continue. Specifically, we examined the number of material enhancements to the iOS and Android operating systems over the period covering the last four major releases from Apple and Google. This includes iOS3 through iOS6 and Android’s Gingerbread, Honeycomb, Ice Cream Sandwich and Jelly Bean releases. We chose these two platforms because, at the time of this writing (Spring 2013), they account for more than 90 percent of the mobile OS smartphone market.



1 New Oxford American Dictionary

2 http://techcrunch.com/2012/11/02/idc-android-market-share-reached-75-worldwide-in-q3-2012/ http://www.idc.com/tracker/showproductinfo.jsp?prod_id=37



While we focused on these two platforms to forecast changes in the OS, we expect continuing innovation in all significant operating systems. Advances in one operating system will fuel accelerated innovation from others.

Our source of data is the documentation that Apple and Google supply developers to inform them of the major additions to the OS in each release.3 Thus we allow the OS vendors themselves to tell us which additions and enhancements to the OS they considered worthy of note. Because, in reality, there are hundreds of changes to the OS in each major release, but only a few broadly indicative of innovation.

The specific items called out by Apple and Google were then placed into one of four architectural layers in the OS, and we counted the number in each layer.

The results of this content analysis of notable enhancements are shown in Figures 1, 2 and 3. Figure 1 combines the data for iOS and Android. There is remarkable consistency in the total number of enhancements to the mobile OS (between 30 and 32) in each new generation, and a fairly consistent number in each layer. The UI and core services layers account for 91 of the 125 or about three-fourths of all enhancements.

3 https://developer.apple.com/library/ios/#documentation/Miscellaneous/Conceptual/iPhoneOSTechOverview/Introduction/Introduction.html#//apple_ref/doc/uid/TP40007898-CH1-SW1 http://developer.android.com/about/index.html

Figure 1: Combined iOS and Android innovation pattern

iOS3 & Gingerbread

iOS4 & Honeycomb

iOS5 & Ice Cream Sandwich

iOS6 & Jelly Bean

Total layer innovations

User interface layer 11 10 12 12 45

Media layer 6 6 7 3 22

Core services layer 11 11 11 13 46

Core OS layer 4 3 2 3 12

Grand total 32 30 32 31 125

Source: Apple Inc. and Google Inc.

Figures 2 and 3 show the same data for iOS and for Android. The first distinctive pattern here is in the total enhancements across all layers of the OS. iOS starts with more, but has trended down over time, concluding with 11 enhancements Apple considered significant enough to signal to the developer community in iOS6. Android starts with fewer, but has trended up concluding with 20 enhancements Google brought to developers’ attention in its Jelly Bean release.

Figure 2: Apple iOS innovation pattern

iOS3 iOS4 iOS5 iOS6 Total layer innovations





Grand total 17 18 14 11 60

Source: Apple Inc.


Both the iOS and Android enhancement patterns show a preponderance of activity in the UI and core services layers. Most of the differences between iOS and Android are explained by an increase in enhancements in the UI layer, from three in the Gingerbread release versus eight in Jelly Bean, its most recent release. Over the same number of generations, iOS has gone from eight UI enhancements in iOS3 to four in iOS6.

(For more details that summarise the actual enhancements we counted for the preceding tables, see our footnote 3 on page 2.)

Figure 3: Android innovation pattern

Gingerbread Honeycomb Ice Cream Sandwich Jelly Bean Total layer

innovations





Grand total 15 12 18 20 65

This view of enhancement patterns over time suggests a forecast pattern for the two currently dominant mobile OSes, and possibly any others.

For example, don’t expect much to happen at the core OS layer, although a few fundamental developments could arise in response to the needs for greater information security, especially in the enterprise context. Beyond that, we are likely to only see improvements required to keep up with incremental advances in the seven enabling technologies—more powerful processing; larger storage on board; higher image resolution, and the like. OS improvements to match those are necessary, not innovative.

Likewise, we do not anticipate multiple major innovations within the media layer. One exception would be a one-time bump in new capabilities as ultra high definition video becomes incorporated in high-end smartphones over the forecast period.

The two layers likely to see the most enhancements over the next five years are the core services and UI layers. Precise predictions are not possible, but we can hypothesise the anticipated vectors of innovation in these layers.

Enhancements in the core services layer will target the developer community directly by exposing advanced hardware features such

as sensor data and making it easier to move between applications and social networks or other cloud access points. Two specific areas likely to see innovation, requiring parallel innovation from the OS, are ubiquitous high-speed connectivity and tighter integration.

Ubiquitous, high-speed connectivity will be essential for many of the proactive always-on monitoring we anticipate from some sensor-driven apps. Next generation wireless starts with the widespread availability of LTE networks that enable the mobile OS to keep the device continuously connected.

But to achieve truly high speed connectivity carriers will introduce hundreds of thousands of self-configuring micro-, pico- and nano-cell base stations, all in service of moving connectivity speeds ever closer to 1 Gibabit (Gb) per second. Only by creating smaller cells connecting fewer users per call will the desired speeds be possible. This introduces additional challenges for the network and the device for managing interference and dealing with multi-cell connectivity.

The LTE environment won’t be everywhere by the end of our forecast period. So the OS will need to be able to figure this out, and find ways to adjust for always-on proactive apps—determining optimal refresh cycles, for example.

Source: Google Inc.


4 http://news.idg.no/cw/art.cfm?id=F75697A1-CE91-D930-2D7D3A8FA64F60DA

As connectivity improves, the OS will need to be able to more thoroughly integrate and orchestrate the mobile device with the cloud, managing the balancing act of native apps on the device with apps, data and services in the cloud and the interplay amongst them; and continuous connection with other devices with which it is federated and has the ability to exchange information. Federated devices might be a desktop computer, a tablet computer and devices belonging to family members or friends with whom the user wishes to interconnect. Or the smartphone might have to federate with wearable computers, acting as the hub and integration point that connects wearable sensors to the cloud.

The desire of users to federate devices could pose a big challenge to OS developers for several reasons. Users tend to have devices from different OEMs, each with its own OS. Sensors are increasingly available with embedded Bluetooth, but standards for data transfer are only beginning to emerge as new sensor applications appear almost daily. While there’s already this kind of integration for a user of all Apple products, the situation is vastly different when the proliferation of intelligent sensors is considered.

We also expect many more enhancements over the next five years in the UI layer. “This is due in part to the opportunity for mobile OS vendors to continue to try to differentiate on the basis of user experiences,” adds Eckert.

The mobile OS has already changed the basic paradigm for how the user interacts with mobile computers, through the capacitive touch sensor just beneath the screen—the swiping, tapping and pinching to manipulate on-screen objects we’re all familiar with.

More recently, voice has become an increasingly important way to literally tell the device what to do. On the horizon is touch-free gesture control, which will require a mobile OS to decide if the movement of a finger was or was not an instruction. We already see wearable computing devices4 that sense the subtle muscle movements associated with the changing positions of individual fingers and translate them into instructions to the OS and any other federated devices.

These are still just variations on how the user can give the smartphone instructions. Through 2015, we expect more instructions to come automatically from sensors and the Web without human intervention. The very definition of ‘user’ and ‘UI’ is likely to expand to include the inanimate pieces of the device that sense, know, and figure things out or scan on behalf of the human but without her intervention.

The mobile OS will acquire the capabilities to orchestrate the services that use the sensors, process the data they capture, and turn it into actionable information. Think of this as making technology ‘people literate’ rather than requiring people to become technology literate.

We already see sensors that act on their own, often with support from some of the standard enabling components on the device (sensor-based apps that work together with the image sensor, for example).

Google Now and its Card metaphor on Android devices are already demonstrating the potential for smartphones to act autonomously on behalf of the user. Early services include tracking the timing and locations of a daily commute by a smartphone owner to ‘learn’ where the user’s home and office are located, then proactively

Figure 4: Patterns of innovation in the OS through 2016

Core layer – Better security and improvements to keep up with incremental advances in the seven enabling technologies.

Media layer – Innovations associated with ultra high definition video.

Core services layer – Enhancements that target the developer community directly by exposing advanced hardware features and making it easier to move between applications and social networks or other cloud access points; specific areas likely to include ubiquitous high-speed connectivity and tighter integration.

User interface layer – Innovations to enhance the growing cognitive capabilities of mobile devices, including more instructions to the device coming from sensors and the web without human intervention.

Source: PwC

© 2013 PwC. All rights reserved. PwC refers to the PwC network and/or one or more of its member firms, each of which is a separate legal entity. Please see http://www.pwc.com/structure for further details. This content is for general information purposes only, and should not be used as a substitute for consultation with professional advisors. BS-13-0351.0413

flagging a traffic jam on the user’s standard route to work, and suggesting alternatives. It is still early days, but the many patterns of behaviour and information interaction in our daily lives offer a plethora of additional ways a digital companion style OS can add value.

Clearly, the evolving mobile OS is beginning to integrate situational information in the service of building a more complete context of the smartphone user, including specific patterns of user behaviour and implied preferences. These broadly come from three sources: the components on the device itself; the surrounding physical environment outside the device as understood by various sensors; and the needs, traits, desires, and preferences of the

user derived algorithmically from an individual user’s application data that builds over time. These cognitive capabilities are also known as contextual awareness.

By contextual awareness we mean that a mobile device understands a user’s relationships to people, places, objects and information, and is able to infer certain needs, intents and goals of the user. Armed with this knowledge, a mobile device can meet a user’s needs and wants with minimal requirement for that person to state them explicitly. In Phase II of the Mobile Innovation Forecast, we will explore contextual awareness in depth, and the role of the OS in providing it.


Raman Chitkara Daniel Eckert Global Technology Industry Leader Mobile Computing Director PricewaterhouseCoopers LLP PricewaterhouseCoopers LLP [email protected] [email protected]


About PwCPwC firms help organisations and individuals create the value they’re looking for. We’re a network of firms in 158 countries with more than 180,000 people who are committed to delivering quality in assurance, tax and advisory services. Tell us what matters to you and find out more by visiting us at http://www.pwc.com


mailto:[email protected]

http://www.pwc.com