Critical communications over mobile operators’ networks ...jultika.oulu.fi/files/nbnfi-fe2019052817560.pdfbroadband networks e.g. long term evolution (LTE) and fifth generation (5G)

2169-3536 (c) 2018 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. Seehttp://www.ieee.org/publications_standards/publications/rights/index.html for more information.

This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI10.1109/ACCESS.2018.2883787, IEEE Access

VOLUME 06, 2018 1

Received November 1, 2018, accepted November 25, 2018. Date of publication xxxx 00, 0000, date of current version xxxx 00, 0000.

Digital Object Identifier 10.1109/ACCESS.2018.Doi Number

Critical communications over mobile operators’networks: 5G use cases enabled by licensedspectrum sharing, network slicing and QoScontrol

Marko Höyhtyä1, (Senior Member, IEEE), Kalle Lähetkangas2, (Student Member, IEEE), JaniSuomalainen1, Mika Hoppari1, Kaisa Kujanpää1, Kien Trung Ngo1, Tero Kippola3, Marjo Heikkilä3,

Harri Posti2, Jari Mäki4, Tapio Savunen5, Ari Hulkkonen6, and Heikki Kokkinen7

1VTT Technical Research Centre of Finland Ltd2University of Oulu, Centre for Wireless Communications3Centria University of Applied Sciences4Airbus Defence and Space5Aalto University, Department of Communications and Networking6Bittium Wireless Ltd7Fairspectrum Oy

Corresponding author: Marko Höyhtyä (e-mail: [email protected]).

This work was supported in part by Business Finland through the CORNET project.

ABSTRACT Commercial mobile operators’ networks will be used for public safety communications dueto demand for wireless broadband services, new applications, and smart devices. Existing dedicatedprofessional mobile radio (PMR) networks, such as terrestrial trunked radio (TETRA), Tetrapol, and project25 (P25), are based on narrowband technologies and hence their data bandwidth is limited. This paperstudies how critical communications needed e.g., by ambulance personnel, rescue squads, and lawenforcement agencies can be implemented over a 5G network. The most important technology enablers aredescribed and test network architectures used in our project given. We focus on two different use cases:First, how to enable priority communications over a commercial mobile network. Second, how to createrapidly deployable networks for emergency and tactical operations. Tests done with the implementedsystems in real networks show that both approaches are very promising for future critical users. Techniquessuch as network slicing and licensed shared access (LSA) provide means to support mission criticalapplications in any environment.

INDEX TERMS Public safety, priority communications, mission-critical communications

I. INTRODUCTIONTraditionally public safety communications services havebeen provided with narrowband professional mobile radio(PMR) systems such as terrestrial trunked radio (TETRA)and Tetrapol in Europe and project 25 (P25) in NorthAmerica [1]. The trend is going towards commercial mobilebroadband networks e.g. long term evolution (LTE) and fifthgeneration (5G) networks in the future due to demands forbroadband services, new applications and smart devices. Forexample, multimedia transmission is useful in many criticalscenarios but current PMR systems support this kind ofservices poorly. The maximum data rate of the TETRA

system is 28.8 kbps and for the P25 system it is 9.6 kbps.Enhanced TETRA may reach a few hundred kilobits persecond which is not yet enough for many multimediaservices. Thus, there is a need for higher transmission rates.There are already ongoing projects for nationwide nextgeneration public safety services based on mobile operatorsnetworks, e.g. FirstNet in the US, Emergency ServicesNetwork (ESN) in the UK and SafeNet in Republic of Korea.Mission critical users such as police officers, border guards,ambulance personnel, and fire and rescue need reliablecommunications, high availability, and security that cannot



VOLUME 06, 2018 9

be matched without numerous technological enablers. Keyenabling technologies for critical communications in fourthgeneration (4G) networks standardized by 3rd generationpartnership project (3GPP) are device-to-device (D2D)communications and proximity services (ProSe), groupcommunications, mission-critical push-to-talk, video anddata (MCPTT, MCVideo, MCData), quality of service (QoS)and prioritization mechanisms including preemption, andend-to-end security.

The situation will be further improved in 5G systems thatwill enable seamless integration of multiple radio andnetwork technologies and even creation of logical publicsafety networks within the same infrastructure as is used forcommercial users [2]. Development of 5G includes a diverseset of technologies such as software-defined networking(SDN), multi-access edge computing (MEC) and spectrumsharing to better support different application areas [3]–[5].Especially licensed spectrum sharing approaches thatguarantee QoS for sharing applications are promising also forrapid deployment of public safety networks.

There is a need for practical testing and trials to verify howthe proposed technologies can be used e.g. in 3GPP evolutionnetworks. 5G test network activities are going on globally, inFinland aim is to build an infrastructure where beyond state-of-the art technologies and flexible service configurations aredeveloped and tested. Recently, we demonstrated citizenbroadband radio service (CBRS) [6] technologies in livetrials with commercial network devices. We haveimplemented and tested licensed shared access (LSA), QoSand policy control mechanisms and network slicingtechniques [7]–[14] to run critical traffic such as MCPTT inour test network. Our trial environment is geographicallydiverse, distributed to several locations in Finland and itsupports multiple domains. In addition, it enables acombination of techniques both on the core network andaccess network sides.

The main motivation for this paper is to show how criticalcommunication use cases are implemented in a 5G network.Our work complements previous studies reported in [1]–[5]by 1) reviewing mission-critical services and latest enablingtechniques and standards in 5G and by 2) implementing andstudying use of commercial technologies for mission-criticalservices in two important use cases. In the prioritycommunication use case we use the policy and charging rulesfunction (PCRF) in the core network for dynamic QoSmanagement to prioritize the MCPTT application. In therapidly deployable network use case, we implement adistributed LTE network for the scenarios where thecommercial network might be unavailable. We further use

LSA and sensing to find the spectrum information for theavailable frequencies for our network.

This article is organized as follows. First we describeneeded technological enablers and review the advanceswithin 3GPP standardization. Then we present the trialenvironment for priority communications and discuss howwe implemented MCPTT application in the network.Another trial environment focusing on a rapidly deployablenetwork and a distributed LSA solution is given in thefollowing part. Finally, we define promising future researchideas and conclude the paper.

II. ENABLERS FOR CRITICAL COMMUNICATIONSMission-critical communication services in mobile operators’networks require multiple technological enablers that arecontinuously advanced in standardization. Various QoScontrol methods can be used to guarantee required level oflow latency, packet error loss, and high priority. We reviewthe status of standardization and describe the mainmechanisms under the 3GPP roadmap. Reliability andintegrity of the communication are crucial for authorities andthus, security mechanisms are needed. In addition,softwarization of future networks and edge computingenables to tailor the network to specific needs of the criticalusers. For example, network slicing can be used to create alogical network inside the physical network to guaranteeenough communication resources e.g. for ambulancecommunications. In some occasions spectrum sharingtechnologies are crucial in finding suitable radio resourcesfor communications e.g. when the current infrastructure hasbeen damaged or lost. Finally, multiple radio technologiesare integrated in the 5G system and they can be seamlesslyused to support different capacity and communication rangeneeds of the users. More detailed description of each enablercan be found in the following subsections.

3GPP roadmap and QoS control: 3GPP is the mainstandardization body for 5G technologies. Figure 1 showshow supporting technologies are included in differentreleases of 3GPP standardization to enable criticalcommunications over commercial networks. The QoSconcept [7] has been included since Rel. 8 and the followingreleases have advanced the QoS mechanisms. Device-to-device communications (D2D) [8] was included already inRel. 12 to enable different proximity services and operationof critical users even without supporting infrastructure.Group communication enablers [9] were developed in thesame release and improved in the following 13th release.Support for mission-critical push-to-talk is included in Rel.13 and for data and video in the Rel. 14.



VOLUME 06, 2018 9

Rel-8-11

5G4G

20142006 2015 2016 2017 2018 2019

Rel-12

Rel-12 equipment

Rel-13Rel-14

Rel-13 equipment

Rel-14 equipment

Proximity,device-to-device& groupcommunications;Quality controlfor mission-criticals

Mission-criticalpush-to-talk;Authenticationand encryptionfor mission-criticals

Mission-criticaldata, video &securityframework;Integration oflicensed sharedaccess controller

2020

Network slicing;Mission-criticalsfor verticals (rails,maritime);Interworking withTetra, P25..,

Rel-15 equipment

Rel-15

QoS parameters;Access barring;Dual-priorityaccess; Policyfunctions; Updatesto 3G pre-emptionservice

Figure 1. Mission critical services and QoS in 3GPP standardization.

In the first phase of 5G system, i.e., in Rel. 15 all themission-critical services are enhanced. In addition,interworking with other systems such as TETRA will be usedso that in the future services could be seamlessly provided tomission critical users with different radio interfaces. The 5Gsystem will include tailored support for different verticalsuch as railway and maritime operations.

QoS control in 3GPP networks ensures that users withhigher priority classification are given access to appropriateoperator resources and receive sufficient service quality evenin congestion situations. Policy-based management i.e.,applying operator-defined rules for resource allocation andresource use, plays a fundamental role in QoS control andtraffic prioritization [7], [12]. QoS in 3GPP networks isbased on the concept of bearers, which is a transmission paththrough the infrastructure and radio interface with a definedcapacity, latency, and packet loss. Bearers are assigned fordifferent applications according to applications’ QoS classidentifiers (QCI) as depicted in Table 1. For each QCI class aresource type that is either guaranteed bit rate (GBR) or non-guaranteed bit rate (NGBR) is given. Every QCI is associatedwith a priority level. Mission-critical services such asMCPTT have the highest priority in the defined QCI classes.

Priority level 5 is the highest priority level and it is assignedfor MC-PTT signaling. Numerical requirements for delayand packet loss rate are shown in the table for each QCI.

Enforcement of QoS policies is based on severalmechanisms. Creation of dedicated bearers and dynamic QoScontrol is enabled with the policy and charging rules function(PCRF) [15]. Access Class Barring (ACB) is mainly used forcongestion control of a specific area or a specific base stationor cell in two ways: 1) Access class control method enablesmobile terminals to determine whether they should send theconnection request to base station; 2) Connection rejectmethod enables base station to accept or reject connectionrequests. Mobile network operators may use both methodsdepending on the network congestion and traffic conditions.Every data bearer must have QCI and allocation andretention priority (ARP) defined. ARP primarily allowsnetwork to decide whether a bearer establishment request canbe accepted or rejected. As priority communication needsmay arise dynamically, 3GPP has defined also preemptionmechanisms that provide means to remove lowest priorityusers so that higher priority users can start immediately theirtransmission when accessing the network.



VOLUME 06, 2018 9

Table 1. QoS class identifiers for different applications [11].Class Resource Type Priority Delay

BudgetError Loss Example Services

1 GBR 20 100 ms 10-2 Conversational voice

2 GBR 40 150 ms 10-3 Conversational (streaming) video

3 GBR 30 50 ms 10-3 Real time gaming

4 GBR 50 300 ms 10-6 Non-Conversational Video

65 GBR 7 75 ms 10-2 Mission Critical user plane Push To Talk voice

66 GBR 20 100 ms 10-2 Non-Mission-Critical user plane Push To Talk

75 GBR 25 50 ms 10-2 Vehicle to everything

5 non-GBR 10 100 ms 10-6 IMS signaling

6 non-GBR 60 300 ms 10-6 Buffered video, TCP-based (www, email…)

7 non-GBR 70 100 ms 10-3 Voice, streaming video, gaming



69 non-GBR 5 60 ms 10-6 Mission Critical delay sensitive signalling

70 non-GBR 55 200 ms 10-6 Mission Critical Data

79 non-GBR 65 50 ms 10-2 Vehicle to everything

80 non-GBR 66 10 ms 10-6 Low latency eMBB, augmented reality

81 Delay Critical GBR 11 5 ms 10-5 Remote control

82 Delay Critical GBR 12 10 ms 10-6 Intelligent transport systems

83 Delay Critical GBR 13 20 ms 10-5 Intelligent transport systems

84 Delay Critical GBR 19 10 ms 10-4 Discrete automation

85 Delay Critical GBR 22 10 ms 10-4 Discrete automation

Priority services rely on 3GPP’s security architecture [16]to authenticate users and to protect network assets againstexternal threats. To support end-to-end confidentiality andintegrity protection, 3GPP has also defined identity and keymanagement mechanisms for mission critical applications[17]. Further, mobile network operators typically applydomain and vendor specific security firewalling andmonitoring solutions to preact and react against disturbancesand attacks affecting the availability of critical services.

Softwarization and network slicing: Promising enablersin 5G networks, affecting a lot both to architectural choicesand mechanisms, include multi-access edge computing(MEC), software-defined networks in (SDN), networkfunction virtualization (NFV), cloud computing and cloudnetworking. They enable deployment and runtime of networkfunctions as software only make networks dynamicprogrammable through centralized control points [2], [15].These technologies are the main drivers for enablingcustomized 5G network infrastructures for specificapplications and services and for seamless integration ofdifferent heterogeneous networks. NFV and SDN can makenetwork slicing (NS) a reality, allowing operators tocustomize networks according to various requirements ofmobile services, thus leading to a more cost-effective way tobuild dedicated networks.

3GPP defines a network slice as a complete logicalnetwork, which provides telecommunication services andnetwork capabilities [13]. Distinct Radio Access Network(RAN) network slices and core network slices will interworkwith each other to provide mobile connectivity. A devicemay access multiple network slices simultaneously through asingle RAN. The realization of NS faces significantchallenges in cellular systems [14], [18]. There are greatdifficulties in organizing mobility and authenticationmanagement in the control plane as well as session andcharging management in the user plane. Network slices areisolated from each other to prevent control plane congestionon one slice to affect the control plane of other slices, and toimprove security.

The simplest way to implement NS would be to createstatic slices for certain purposes. However, in criticalcommunications the demands are dynamic and thus, dynamicnetwork slicing is a preferred way of operation1. Thefollowing phases describe the network slice lifecycle [13]: 1)Preparation phase, 2) Instantiation, Configuration andActivation phase, 3) Run-time phase, 4) Decommissioning

1 However, an always available dedicated slice in commercial networksfor critical communication purposes could be very useful. The drawbackfor this approach is resource consumption since the slice might beunderused majority of time.



VOLUME 06, 2018 9

phase. Run-time monitoring and controls enable dynamicmanagement the functionality and resources of slice.Network slices can be dedicated or shared among multipleapplications or users [18]. The slices may have dedicated orcommon network functions. Dedicated functions facilitatecustomization of slices and assure strong isolation betweendifferent applications. A fixed splitting of common functionsand resources simplifies the network management andoperation but may lead to inefficient network utilization.

Spectrum sharing and multi-radio access technologies(RATs): Spectrum sharing approaches support criticalcommunications by enabling use of more spectrum resources[19]. Typically, critical communications are the primary userin the sharing arrangements. Allowing commercial and otherspectrum users on the critical communications bands canreduce the risk of being migrated to another band, even if theuse of the radio spectrum is rare. Organizations utilizingcritical communications may also benefit from secondaryaccess to commercial spectrum bands, for example fortraining purposes. Spectrum sharing with a primary statusbetween different critical communications systems could alsobe a possible scenario in the Public Protection and DisasterRecovery (PPDR) missions. Distributed LSA is a promisingapproach for rapidly deployable networks. The basicprinciple of a spectrum database approach is that the databasecontrols when the secondary user can access the spectrumand such helps to avoid harmful interference to protectedspectrum users. Under the LSA approach, the incumbentoperators are required to provide information about theirspectrum use over the area of interest to the database. Basedon that information, the spectrum database determinesexplicitly where, when, and which parts of the frequencybands are available for the secondary use. This approachenables interference free spectrum sharing for all andguaranteed QoS for the incumbent networks.

5G is a multi-radio system built upon both new highcapacity and low-latency interfaces and convergence ofexisting radio technologies such as 5G new radio (NR), LTEand WiFi to a ubiquitous radio access network. The wirelessaccess can be obtained with multi-operator 3G/4G/LTErouter. The routers can be combined with a high gaindirectional narrow beam antenna high of the ground whenlong range communications are required. This setup providesconnection in areas where the normal LTE mobile phones donot have coverage. Then, for even more remoteenvironments, satellite broadband connections are availablein integrated satellite-terrestrial 5G networks [20]. Thismulti-radio network requires intelligent end-to-end networkmanagement where selection of the most appropriate RATand route to the data is based on the QoS requirements andmeasurements.

III. CRITICAL COMMUNICATIONS TRIALSOur practical work has focused on two use cases. First, howto implement critical communication services in the existing

mobile operator network using described technologicalenablers? We focus on push-to-talk application in Trial 1,describing how we have implemented the system in our testnetwork and what are the main components andcommunication interfaces between those components. Trial 2focuses on creation of a network in an area whereinfrastructure is not currently available. The rapidlydeployable network (RDN) requires that all the requiredfunctionalities such as core network, application servers, andthe LSA server for spectrum needs are implemented and builtin the area on interest. Finally, collaboration trials betweenthe RDN and the commercial 5G are presented.

A. PRIORITY COMMUNICATIONS OVER MOBILEOPERATORS’ NETWORKSThis use case focuses on enabling critical communication incommercial networks by use of QoS management andprioritization mechanisms. We are using dedicated bearersand mission-critical QCI classes for those bearers toprioritize PTT users in the commercial networks. This can beseen as “4G network slicing” that we were able to implementalready with existing network components. In the future 5Gnetworks the slicing will be based on software-definednetwork switches and network function virtualizationmechanisms. The proposed “4G network slicing” withdedicated bearers could be implemented in the existingcommercial networks already now. Figure 2 describes theimplemented trial environment for priority communications.The key building blocks and interfaces of the trialenvironment are given in the following. The trial architectureis geographically distributed in several locations in Finland.The architecture connects the application server in Helsinkiarea to the core network in Oulu and can be used to supportPTT clients over an access network both in Helsinki and inOulu. The architecture enables adding flexibly componentsand testing technologies in different locations andenvironments.

PTT application: One or more PTT clients requirepriority connection in the commercial network for authoritycommunications, i.e., using pull mode from the end -users’side to open the connection. The PTT client sends servicerequest over the LTE-Advanced network to the PTTapplication server.

PTT application server provides the groupcommunication and messaging service to the PTT clients. Italso interfaces to the narrowband TETRA networksproviding seamless interoperability between PTT clients andTETRA radios. The server uses Rx interface to communicatewith the PCRF located in the core network and requestspriority connection to the PTT clients based on a servicerequest, PTT client user profile, and service profileprovisioned in the server.



VOLUME 06, 2018 9

X2

UE 3

UE 2:PTTclient

eNB 2

eNB 1

PTT application server

MME HSS/SPR

PCRF

S6a

SGW

S1-U

S11

Public safety serviceprovider applications

SGi

Rx

Access network Core network

PGW

Uu S1-MME

Gxx

Gxx

Uu

Uu

UE 1:PTTclient

Low priorityusers, e.g.

entertainment

High priorityusers

Figure 2. Test network architecture for priority communications.

Figure 3. Throughput of different priority users in the trial network.

Core network is implemented with the OpenEPC [21].The main components regarding the priority communicationare given in the figure. A key component is the PCRF thatprovides QoS of different applications and subscribers. It canalso apply security procedures before accepting informationfrom the PTT application server that is under control of thepublic safety service provider. The PCRF is used in creatinga dedicated bearer for the PTT application with a missioncritical QCI 65 value. The PCRF allows dynamic creationand management of dedicated bearers, i.e. new bearers canbe added on the fly to the network.

Other main component of the core network is the mobilitymanagement entity (MME) that is the main signaling node inthe evolved packet core (EPC), taking care of e.g.authentication and handover signaling between differentnetworks. MME connects to the base station, evolved node B(eNB), through the S1-MME interface and connects to theserving gateway (SGW) through the S11 interface. MMEworks with the home subscribing server (HSS) and the RANto decide the appropriate radio resource management strategythat can be UE-specific. The HSS includes user identificationand addressing data as well as subscriber profile repository(SPR) where user-subscribed QoS classes are stored. ThePCRF is connected to the SGW, packet data network (PDN)gateway (PGW), HSS and MME using Diameter based

connectors Gxx, Gx, and Sp. The PDN Gateway providesconnectivity from the UE to external packet data networksand public safety applications over the SGi interface. The S5reference point provides tunneling and management betweenthe SGW and the PGW.

Access network: The used RAN in our test network islocated in office environment. It includes 3GPP base stationsthat are connected to each other using X2 interface. The basestations and UEs are commercially available 3GPP compliantequipment that are constantly updated with the latestsoftware releases.

We have tested creation of dedicated bearers and use ofthem in a congested network, especially the mission-criticalapplications. The results shown in Figure 3 were achievedwith our trial network using a pico base station and three UEswith different QCI class bearers. First, there is a 30 Mbits/sQCI9 default bearer data transmission to UE3 (green line)going on, this user being the only one over the air. Thetransmission stops after data transmission to the user UE1(red line) with a high priority dedicated bearer QCI5 starts.The result shows that the priority user reserves the wholeband for its use. After another critical user UE2 (blue line)with the same priority and dedicated bearers QCI5 startstransmission UE1 and UE2 divide the bandwidth equally



VOLUME 06, 2018 9

Public safetyoperator

LTE access

point

LTE access

point

4G/LTE/Bridge/APN/LSA server

LSA repository

Incumbent(Commercial

mobile operatornetwork)

LSA spectrumwhere, when

LSA spectrumavailability

NationalregulatoryauthorityLSA license

Datasharingpolicy

LSA licenseand sharingframeworknegotiation

LSA system

Tacticalrouter

Radiohead

Tacticalrouter

Radiohead

Lite-EPC

DistributedLSA and

spectrumcontroller

Lite-EPC

DistributedLSA and

spectrumcontroller

S1

S1

Figure 4. A concept for rapidly deployable public safetynetwork using LSA for spectrum needs.

while they are receiving data simultaneously. Traffic to UE3default bearer continues after UE1 and UE2 have stopped.The way to use the bandwidth can be affected by policymanagement in the core network. For example, in our casethe priority users may request total bandwidth for their usebut there could be also a policy set to leave e.g. 20 % of thebandwidth available for commercial users. This could be stilllimited so that video transmission is not allowed but e.g.speech and text is fine.

This trial shows how critical communication users canreceive a high quality service in a commercial network whenneeded. The trial scenario enabled us to evaluate the maturitylevels of available commercial products. For instance, wenoticed that the support for mission-critical quality classeswas inadequate in the base station and consequently weutilized lower quality classes with identical behavior. Wewere also able to advance development of the EPC to supportbetter GBR bearers.

B. RAPIDLY DEPLOYABLE NETWORKSThis use case focuses on building a rapidly deployablenetwork (RDN) for public safety applications for scenarioswhere the commercial network is fully or partiallyunavailable. This could be caused by some catastrophicsituation such as an earthquake or there might be a need tocreate a dedicated network in a remote location for searchand rescue, as an example. The rapidly deployable networkprovides the basic services such as voice and data transfer forthe users as well as multiple ways to access the network andits services including local LTE, WiFi and access via acommercial mobile network, if available. Here, drones,sensors, and cameras can be connected to the networkwirelessly or wired. Figure 4 describes the rapidly deployablenetwork concept including the LSA functionality that is used

Figure 5. An example public safety backhaul network.

for spectrum management [4]. The RDN can utilize towablelifts or crane cars to allow the rapid deployment andsufficient radio coverage without fixed tower assemblies.The test environment is Finnish forest.

The lite-EPCs provide LTE access points to the tacticalnetwork and emulate the EPC functionalities of a commercialLTE network. The lite-EPC is a server program runninginside a small and portable computer. Compared with thecomplete commercial EPCs, this is a distributed and a lightersolution to enable stand-alone LTE operation for the selectedusers.

The tactical network offers the backhaul for the publicsafety operator network and consists of SDR-routers [22],which support both wireless and wired connections. Theserouters are able to select and use the best possibleconnection/route available and will re-route the connectionautomatically, if needed. In addition to tactical connections, aconnection to any public and mobile networks can beprovided allowing the users connect to, for example, to theinternet, or remote users connected to a mobile network toaccess the tactical network through a secure Mobile virtualprivate network (VPN) connection. Note, that while in Figure4 we have two tactical routers, the network is scalable, i.e.,the number of the access points and tactical routers can vary.As an example, Figure 5 shows a tactical network trial withfour backhaul nodes. We also see a crane car that has all thenecessary equipment. The trial included backbone links up to12 km distances, which provided sufficient connectivity ateach location and enabled group communication betweenLTE handhelds.

The LSA system offers spectrum information for therapidly deployed LTE network. A secondary spectrumlicense is sufficient, because this network is a backup



VOLUME 06, 2018 9

Figure 6. Rapid network deployment with crane cars.

network for the commercial network. Here, the incumbentspectrum users reserve the spectrum an agreed time beforetheir transmission. The LSA repository saves thisinformation. The LSA server obtains the latest spectruminformation from the repository, which is then valid for theagreed time after the LSA server has had a repositoryconnection. The LSA server acts as a mediator between theLSA repository and the distributed LSA controllers.

The distributed LSA controllers are running on the samecomputers as the lite-EPCs. The controllers control their ownLTE access point spectrum use, but decide the spectrumsharing together. They synchronize with each other andobtain spectrum information from the LSA repository andfrom the sensing system. This information enables to controlthe access point for locking/unlocking cells, changingfrequency, and/or changing transmission power. Thecontroller is a small and lightweight software that providesrapid base station control. It is easy to port to any computerand thus easy to utilize in mobile scenarios. The performedtrials verified the LSA functionality for this type of rapidlydeployed network operation.

The sensing offers additional spectrum information for therapidly deployed network. It is used if the LSA informationis uncertain or invalid due to commercial network breakage;the sensing is used both to verify the spectrum informationvalidity and to select the channel with the least interferencewhen the LSA system fails. In our trials, a spectrum analyzerwas used remotely to calculate the channel energy ofdownlink LTE Band 7. Initial tests suggest that sensing theLTE channel free with an energy detector and by using

antenna gain 6 dBi and antenna height 30 meters canguarantee sufficient separation distance to the incumbentmacro base station in remote areas. We plan to use theradioheads for sensing.

To understand the dynamic spectrum use capability of ourtrial system, we performed measurements with thedistributed LSA setup while evacuating the frequency andstarting the transmission in a new band. The evacuationprocess aims to protect the incumbent spectrum users’ rightsby evacuating LSA band when the band request from theincumbent is received. The restart frequency process on theother hand defines the necessary time for a base station tostart their service in a new LSA band from the time the basestation received a new frequency band information from theLSA repository. Thus, we measured time consumption inboth processes in order to show the efficiency of own newLSA concept in this experiment.

The measurement begins at the moment the LTE basestation received the information from the LSA repository thatthe current frequency is not available anymore. Followingsteps are then performed: 1) At the base station, LSAcontroller executes LOCK procedure to turn off the airinterface at the current frequency. 2) After that, the LSAcontroller inquires new information about available spectrumfrom the LSA repository and synchronizes the newest updatewith other LSA controllers in the network. 3) Then, thecontroller uses new information to find available frequencyand executes SET_FREQUENCY procedure to reconfigurethe base station to operate in a new spectrum band. 4) Finallycontroller executes UNLOCK procedure to turn the airinterface of the base station back on to continue transmissionon the configured frequency.

The results are shown in Table 2. Our previousmeasurements that have used large network managementsystems [6] instead of our self-developed controllers haveshown reconfiguration times in the order of minutes. Now weachieved much better results for a rapidly deployablenetwork where we are only controlling the small networkwith dedicated controller functionality. The achieved timesshow that current band can be evacuated in less than 10seconds during LOCK command. The restart frequencyprocess combining the SET FREQUENCY and UNLOCKtimes takes approximately 17 seconds.

TABLE 2. Reconfiguration times in the rapidlydeployable network trial.

System Commands Timeduration

StandardDeviation

LOCK - Turn off air interface at basestation

8.78 s 0.39 s

SET_FREQUENCY - Change LSAfrequency from one to another

7.97 s 0.37 s

UNLOCK - Turn on air interface atbase station

8.97 s 0.37 s



VOLUME 06, 2018 9

C. COLLABORATION BETWEEN COMMERCIAL ANDRAPIDLY DEPLOYABLE PUBLIC SAFETY NETWORKSIn addition to the conducted main trials and achieved results,we have tested the RDN concept collaboratively withcommercial networks. The tests have been conducted duringpublic safety operations so that the personnel has used morethan 20 end user equipment that can access both RDN andcommercial network. The solution has ensured operationalsafety and priority communications in commercial network.Based on the operational tests following requirements andrecommendations can be given.

The collaboration between commercial and public safetynetworks can be enabled with a political decision.Alternatively, the commercial operators can have monetaryincentives to offer tailored and complete solutions for publicsafety together with the rapidly deployable networkequipment providers. The public safety needs interconnectionand requires a similar operation of the applications andservices in every network. The public safety can utilizecommon applications in each type of network in thefollowing fashion.

Common services at the internet: The public safetyoffers and uses its application servers and data bases over theinternet. The services can be hidden services in the tacticalnetworks and alternatively offered in any other private orpublic networks. These applications are reachable from anynetwork having an internet connection. The connection canbe secured for example with a VPN. Normally, the publicsafety can gain wireless internet with single or multi-radioaccess technologies and in planned operations, public safetycan use locally available commercial hardwired networks.

Common services at the commercial network EPC: Thecommercial network configures the public safety basestations at the tactical networks to its EPC. Likewise, thepublic safety configures its base stations to use thecommercial EPC. The main benefits of this scenario are thelatest commercial application services and theinteroperability with older mobile systems.

The public safety can connect to the EPC via a VPN usingan internet connection or via commercial network slices.Then, when there is no connection, the public safety canutilize a local EPC solution, which can support the servicesthat the public safety has noticed to be beneficial.

Note that the public safety can also act as a mobile virtualnetwork operator inside the commercial networks.Furthermore, the public safety can operate a full commercialEPC service itself. Here, the public safety obtains morecontrol over the network it uses. Then, it can have roamingagreements for the communication with the commercialnetworks. With roaming, the public safety can allowcommercial users to use their base stations. This can becritical in some operational scenarios, where the commercialnetwork is down and the commercial users do not have anoutside world connection.

Common commercial network switches and routers:The commercial networks can enable the backbone of thecommercial networks for public safety use. Here, any of theconnected networks can offer the applications, which areaccessible to critical users via routing. Rapidly deployedpublic safety networks are connected to commercialnetworks using interface devices, which are secure trafficrelays. The public safety can have multiple interface devicesin different parts of the commercial internet protocol (IP)networks. They can be connected to each other with methodslike a virtual local area network (VLAN) or a virtual privatelocal area network service (VPLS). Then, they cancommunicate together with their desired method. Note thatthese interface setups can be dynamically set up or pre-defined permanent installations.

The interface devices can also give tactical network accessto the LTE users outside of the public safety operatornetwork. Here, the interface device has been defined a knownaccess point name (APN) for critical data. Moreover, thecritical LTE user equipment request the same APN in theircritical applications. Note that the LTE has support formultiple packet data networks [23]. Thus, the non-criticalapplications can route their data to elsewhere.

IV. FUTURE RESEARCH TOPICSDynamic network slicing: Network slicing can improvemany public safety services by decreasing delays. Forexample, QCIs are currently requested on per session basispotentially slowing down overall call control setup. Alreadya static network slice enables reserving the resources aheadof time in a coarse-grained manner instead of per session.Since static slicing consumes resources, dynamic slicingtechniques will be needed in future networks to quicklycreate, adapt, and manage slices according to the needs ofusers and applications. Dynamic slicing provides alsorobustness against challenging network conditions, asapplication specific network slices can be more easilyisolated from other applications and adapted e.g. in the caseof emerging cyber security attacks or hardware failures. Itis an open research challenge how to do these in practice.

Softwarization and virtualization of the networkresources are key enablers for the dynamicity [24]. One ofthe core challenges currently is to identify or develop a setof technologies suitable to implement the infrastructureover which network slicing will be built, without requiringmajor rework of the 3GPP specifications. In the longerterm, trials of dynamic network slices for prioritycommunications between rapidly deployed networks is apotential topic to be tested in our trial environment.Traditional QoS and prioritization mechanisms will be usedin network slices to prevent congestions and emergingpossibility to dynamically adjust e.g., the size of the slicewill further improve the situation. Open research challengesinclude how to apply machine learning based controlstrategies in dynamic adjustments.



VOLUME 06, 2018 9

Remote operations: Majority of work conducted withprioritized public safety communications has been done indensely populated areas and due to congestion probabilitiesit will remain as the most important environment also in thefuture. However, prioritized PTT applications incommercial and in rapidly deployable networks in remotelocations is a topic that requires more work to be done.How to create rapidly deployable networks with D2Dconnections and slices in those areas? How to efficientlyuse multi-RAT networks, combining satellite systems,long-range digital communications and broadband solutionsin the remote and isolated environments to differentservices? QoS-aware radio and network resourcemanagement schemes will be required. Partly this work isdone by developing solutions for different mobile platformssuch as cars [20], [25] ships [20], unmanned aerial vehicles[26] and trains [27].

V. CONCLUSIONS

This paper discusses how critical communications can becarried out in a mobile operators’ network and how adedicated rapidly deployable network can be implementedto support public safety needs. The most important enablingtechniques such as prioritization, QoS management, SDN,network slicing, end-to-end security, and spectrum sharingare discussed and their use in practical trials described.Architectures of trial networks used for prioritycommunications and rapidly deployable operations aregiven. The results shown in each trial indicate that thedeveloped mechanisms in prioritization and spectrumsharing are good enablers for future public safety users. Thepaper defines promising future research questions,especially more trials are needed to validate technologies.

VI. REFERENCES

[1] A. Kumbhar, F. Koohifar, I. Guvenic, and B. Mueller, “Asurvey on legacy and emerging techonologies for publicsafety communications,” IEEE Commun. Surveys Tuts., vol.19, pp. 97–124, 1st Quart., 2017.

[2] C. Sexton et al., “5G: Adaptable networks enabled byversatile radio access technologies,” IEEE Commun. SurveysTuts., vol. 19, pp. 688–720, 2nd Quart., 2017.

[3] M. Höyhtyä et al., “Spectrum occupancy measurements:Survey and use of interference maps,” IEEE Commun.Surveys Tuts, vol. 18, pp. 2386–2414, 4th Quart., 2016.

[4] R. H. Tehrani, S. Vahid, D. Triantafyllopolou, H. Lee, and K.Moessner, “Licensed spectrum sharing schemes for mobileoperators: A survey and outlook,” IEEE Commun. SurveysTuts., vol. 18, pp. 2591–2623, 4th Quart., 2016.

[5] O. Ergul, G. A. Shah, B. Canberk, and O. B. Akan,“Adaptive and cognitive communications architecture fornext-generation PPDR systems,” IEEE Commun. Mag., vol.54, pp. 92–100, Apr. 2016.

[6] M. Palola et al., “Field trial of the 3.5 GHz citizensbroadband radio service governed by a spectrum accesssystem (SAS),” in Proc. DySPAN, Mar. 2017.

[7] 3GPP TS 23.107 V15.0. “Quality of Service (QoS) conceptand architecture.” 2018.

[8] 3GPP TS 36.877 V12.0. “LTE Device to Device ProximityServices.” 2015.

[9] 3GPP TS 23.468 V15.0. “Group Communication SystemEnablers for LTE.” 2017

[10] M. Alasti, B. Neekzad, J. Hui, and R. Vannithamby, “Qualityof service in WiMAX and LTE networks,” IEEE Commun.Mag., vol. 48, pp. 104–111, May 2010.

[11] 3GPP TS 23.501 V15.2. “System Architecture for the 5GSystem.” 2018.

[12] 3GPP TS 29.214 V14.3.0 (2017-03), “TechnicalSpecification Group Core Network and Terminals; Policyand Charging Control over Rx reference point (Release 14).”

[13] 3GPP TR 28.801 V1.2.0 (2017-05), “Study on managementand orchestration of network slicing for next generationnetwork (Release 15).”

[14] X. Zhou, R. Li, T. Chen, & H. Zhang, “Network slicing as aservice: enabling enterprises' own software-defined cellularnetworks”, IEEE Commun. Mag., vol. pp. 146–153, July2016.

[15] H. Nam, D. Calin, H. Schulzrinne, “Intelligent ContentDelivery over Wireless via SDN”, in Proc. IEEE WCNC,Mar. 2015.

[16] 3GPP TS 33.501 “Security architecture and procedures for5G system (Release 15),” 2018.

[17] 3GPP TS 33.180 “Security of the mission critical service.”2018.

[18] P. Rost et al., “Network slicing to enable scalability andflexibility in 5G mobile networks,” IEEE Commun. Mag.,vol. 55, pp. 72–79, May 2017.

[19] M. M. Sohul, M. Yao, X. Ma, E. Y. Imana, V. Marojevic,and, J. H. Reed, “Next generation public safety networks: Aspectrum sharing approach,” IEEE Commun. Mag., vol. 54,pp. 30–36, Mar. 2016.

[20] M. Höyhtyä, T. Ojanperä, J. Mäkelä, S. Ruponen, and P.Järvensivu, “Integrated 5G satellite-terrestrial systems: Usecases for road safety and autonomous ships,” in Proc.KaConf, Oct. 2017.

[21] M. Corici et al., “OpenEPC: A technical infrastructure forearly prototyping of NGMN testbeds,” in Proc. TridentCom,pp. 166–175, May 2010.

[22] Elektrobit “Enhancing the link network performance with EBtactical wireless IP network (TAC WIN),” in EB DefenseNewsletter, Dec. 2014.

[23] ETSI, “LTE; general packet radio service (GPRS)enhancements for evolved universal terrestrial radio accessnetwork (E-UTRAN) access,” in ETSI TS 123 401 V14.6.0,Jan 2018.

[24] R. Mijumbi, J. Serrat, J.-L. Gorricho, N. Bouten, F. D. Turck,and R. Boutaba, “Network function virtualization: State-of-the art and research challenges,” IEEE Commun. SurveysTuts., vol. 18, pp. 236–262, 1st Quart. 2016.

[25] W. Siun, D. Yuan, E. G. Ström, and F. Brännström, “Cluster-based radio resource management for D2D supported safety-critical V2X communications,” IEEE Trans. Wirel.Commun., vol. 15, pp. 2756–2769, Apr. 2016.

[26] A. Merwaday and I. Guvenic, “UAV assisted heterogeneousnetworks for public safety communications,” in Proc.WCNCW, Mar. 2015.

[27] A. Sniady, J. Soler, M. Kassab, and M. Berbineau, “Ensuringlong-term data integrity: ETCS data integrity requirementscan be fulfilled even under unfavorable conditions with theproper LTE mechanisms,” IEEE Veh. Technol. Mag., vol. 11,pp. 60–70, June 2016.

Documents

Critical communications over mobile operators’ networks ...jultika.oulu.fi/files/nbnfi-fe2019052817560.pdfbroadband networks e.g. long term evolution (LTE) and fifth generation (5G)