Upload
dokien
View
223
Download
0
Embed Size (px)
Citation preview
Jong T. ParkKyungpook National University
Management of Management of
Ubiquitous Sensor NetworkUbiquitous Sensor Network
2APNOMS 2005 Tutorial, Okinawa, Japan
Outline
Brief Overview on Ubiquitous Sensor Network TechnologyUbiquitous Sensor Network and its ApplicationsUSN Characteristics and Research Projects Network Protocol Architecture of USN Current Standardization ActivitiesLocation and Synchronization Technology in Sensor Network
USN Management RequirementsUSN Management Architecture
USN Management ArchitectureUSN Management Functional AreasRelated Research on USN Management Architecture
Specific USN Management FunctionsPower ManagementTopology ManagementSecurity ManagementContext Management
Conclusion
3APNOMS 2005 Tutorial, Okinawa, Japan
What is Ubiquitous Sensor Network?
Infrastructure network for realizing ubiquitous computing environment using may sensor nodes with sensing,processing & wireless communication capabilitiesCheap and smart sensor node deployed, and monitoringand controlling target environment
Internet Sink
Sensor ADCProcessor
StorageTransceiver
Sensing unit Processing unit Transmission unit
Target
Sensor node
User
Power UnitPower
Generator
Position finding system Mobilizer
4APNOMS 2005 Tutorial, Okinawa, Japan
Sensor Networks
Sensor network is composed of a large number of sensor nodesSensor nodes are small, low cost, low-power devices that have following functionality:
Communication on short distances due to power limitationSense environment dataPerform limited data processing
Network usually also contains “sink” node which connects it to the outside world
5APNOMS 2005 Tutorial, Okinawa, Japan
Sensor Node Hardware
19.1 Kbps 20m Range Light Sensing Temperature Sensing 4 MHz – 3.0 V 8 Kbytes –Program Memory 512 Bytes – Data Memory Available from: CrossBow Inc. $900 for a complete kit
UC Berkeley Motes
MIT µAMPS-I node
6APNOMS 2005 Tutorial, Okinawa, Japan
Sensor Network Applications
Military ApplicationsEnvironment and Habitat MonitoringManufacturingTransportation Seismic StudyHealth Care Home Network
7APNOMS 2005 Tutorial, Okinawa, Japan
Sensor Network Applications (Cont.)
Military: battlefield surveillance, enemy/friendly forces monitoring & tracking, biological attack detection, battle damage assessmentEnvironmental and Habitat : file detection, flood detection, agricultural usesHabitat monitoringManufacturing: inventory controlTransportation: traffic control, shipping and commodity managementHealth: human physiological data monitoringMiscellaneous: car theft detection
8APNOMS 2005 Tutorial, Okinawa, Japan
Characteristics of USNHardware Constraints
Low cost (wasteful): temperature, lightning condition, object presence and movement, pressure etc.Low power, and limited processing/memory capacities: sensing node, sink node, and manager node
High DensityLarge number of sensor nodes (maybe 10 to 100,000 nodes): Scalability problemNode position may not be predetermined
Low Energy ConsumptionLife time of sensor network depends on battery life timeTo relocate & recharge large number of sensing nodes is impossibleNeed adaptable power management strategy: multi-hop communication mode, energy-conservative applications
Network Self-configurationLarge number of nodes in hostile locations-> manual configuration unfeasibleNodes may fail & new nodes join the network: node failures should be regarded as normal conditionAd-hoc sensor network protocols may be applied
Task-specific Application and Data-centric routingNetwork application is defined and developed specifically for each sensing taskData-centric routing is usually employed so that attribute (or location)-based naming are used.
Data fusion and query for collaborative/distributed processingLocally carry out simple computation -> forwards and aggregate dataQuery for single node or group of nodes based on attribute and/or locationBase nodes collect data from given area & create summary messages
9APNOMS 2005 Tutorial, Okinawa, Japan
USN and Ad-Hoc Network Comparison
Global ID(IP Address)No Global IDID
Node IDAttribute/Location-BasedAddressing/Naming
Relatively highLimitedResource (Processing)
Almost steadyVery frequentTopology Change
Not Prone to FailureProne to failureFailure
RechargeableVery Limited Power Consumption
Point-to-pointMulti-hop, Power-conservative Broadcast/multicast,
Communication mode
Relatively sparselyDensely (about 20/meter cube)Deployment
10 ~ 100100 ~ 1000Number of Nodes
Ad-hoc NetworkSensor NetworkItems for Comparison
Needs new paradigm on communication protocols and its MANAGEMENT !
10APNOMS 2005 Tutorial, Okinawa, Japan
USN Research Projects
SMART DUST – UC Berkeley•Autonomous Node 1mm3, MEMS Tech.•Optical Comm. Module, CCR•Sensing, Processing, Communication•Small Size, Low Power, Low Cost
SMART DUSTSMART DUST – UC Berkeley•Autonomous Node 1mm3, MEMS Tech.•Optical Comm. Module, CCR•Sensing, Processing, Communication•Small Size, Low Power, Low Cost
Things The Think (TTT) - MIT Media Lab.•Embedding computation into everyday things. (such as clothing, jewelry, and tables.)•Large amount of Project.(Wearable Health, Smart City, UbER-Badge etc.)
Things The Think (TTT)Things The Think (TTT) - MIT Media Lab.•Embedding computation into everyday things. (such as clothing, jewelry, and tables.)•Large amount of Project.(Wearable Health, Smart City, UbER-Badge etc.)
Smart Kindergarten – NESL-UCLA•iBadge, Childhood education environment •Monitoring & AnalysisEvaluations of students progress“How well is student A reading the story book B?”
Smart Kindergarten Smart Kindergarten – NESL-UCLA•iBadge, Childhood education environment •Monitoring & AnalysisEvaluations of students progress“How well is student A reading the story book B?”
TRON (The Real-time OS Nucleus) - Japan•ITRON(Industry TRON), BTRON(Business TRON)•T-Engine (2002. 6)
- Tron based Development platform- 400 corporations (2004) : MS, Samsung
TRON (The RealTRON (The Real--time OS Nucleus) time OS Nucleus) -- JapanJapan•ITRON(Industry TRON), BTRON(Business TRON)•T-Engine (2002. 6)
- Tron based Development platform- 400 corporations (2004) : MS, Samsung
11APNOMS 2005 Tutorial, Okinawa, Japan
USN Research Projects (cont’)
Oxygen – MIT Computer Science Lab• Computing service available like Oxygen• Computing access using human centered interface such as
natural language and vision
OxygenOxygen – MIT Computer Science Lab• Computing service available like Oxygen• Computing access using human centered interface such as
natural language and vision
CoolTown – HP• Real World Wide Web• Person, place or thing existing on the Web space,
communicating each other• Education, Medicare, ITS, Fire and Safety Service
CoolTownCoolTown – HP• Real World Wide Web• Person, place or thing existing on the Web space,
communicating each other• Education, Medicare, ITS, Fire and Safety Service
SMART TAG – MIT Auto-ID Center•RFID, Supply Chain Management•Collaboration with EPC Global•EPC Code, ONS, PML•Low Cost TAG (Cost < 5 Cent)
SMART TAGSMART TAG – MIT Auto-ID Center•RFID, Supply Chain Management•Collaboration with EPC Global•EPC Code, ONS, PML•Low Cost TAG (Cost < 5 Cent)
Global Supply Chain
Easy Living Project – Microsoft•Intelligent Environment•Info. Gathering Tradition I/O Device – Mouse, Keyboard, MICSensing Device – Cameras, Active Badge
Easy Living ProjectEasy Living Project – Microsoft•Intelligent Environment•Info. Gathering Tradition I/O Device – Mouse, Keyboard, MICSensing Device – Cameras, Active Badge
12APNOMS 2005 Tutorial, Okinawa, Japan
USN Research Projects (cont’)
u-Korea – Ministry of Information and Communication• Development of Key Technologies in Wireless Personal Area
Network, UWB (~hundred Mbps), Electronic Tag (RFID), Intelligent Wireless Sensor Network
• Ubiquitous Home Network and Commodity Circulation Network
uu--KoreaKorea – Ministry of Information and Communication• Development of Key Technologies in Wireless Personal Area
Network, UWB (~hundred Mbps), Electronic Tag (RFID), Intelligent Wireless Sensor Network
• Ubiquitous Home Network and Commodity Circulation Network
Smart-Its Project – ETH, TecO, VTT•Disappearing Computer Initiative (16Projects)•Development Device (Disappearing Computer Initiative)
•Smart Cup (Temp. & User Detecting)
SmartSmart--Its ProjectIts Project – ETH, TecO, VTT•Disappearing Computer Initiative (16Projects)•Development Device (Disappearing Computer Initiative)
•Smart Cup (Temp. & User Detecting)
U-Network – Ministry of Internal Affairs & Communications•Net. Roaming Based On Small Chip & Terminal•Anywhere Connection•Hotspot Net. Service Mobile Network Service•Project: TRON (Tokyo UNIV.) eHII (Matsushita)
UU--NetworkNetwork – Ministry of Internal Affairs & Communications•Net. Roaming Based On Small Chip & Terminal•Anywhere Connection•Hotspot Net. Service Mobile Network Service•Project: TRON (Tokyo UNIV.) eHII (Matsushita)
Network Protocol Architecture of USN
14APNOMS 2005 Tutorial, Okinawa, Japan
Sensor Network Protocol Architecture(Akyildiz, etc, GIT)
Pow
er man
agemen
t plane
Mobility m
anagemen
t plane
Task man
agemen
t plane
Application layer
Transport layer
Network layer
Data link layer
Physical layer
Application specific protocols
Frequency selection, carrier frequency generation, signal detection, modulation, data encryption
Schedule the sensing tasks given to a specific region and/or power level
Movement detection, neighbor detection and registration
Manage how a sensor node uses its power
Needed when access through the Internet or other external networks
Multihop wireless routing protocols between the sensor nodes and the sink nodes
Multiplexing of data streams, data frame detection, medium access and error control
15APNOMS 2005 Tutorial, Okinawa, Japan
Logical Functions of USN Protocol Layers
Sensor Query, data dissemination/aggregation,
Task assignment, Management
Maintain connections to external Internet: TCP splitting
Data-centric routing, Self-configuration, Geo-Routing
MAC, Time and Location, Adaptive
Communication, Sensing, Actuation
Physical
Data Link
Network
Transport
ApplicationCoordinated to minimize power consumption
Adaptive topologyEnergy-aware RoutingAdaptive MAC
In-network processingData-centric routingMulti-hop communicationNo direct communication toservers
16APNOMS 2005 Tutorial, Okinawa, Japan
Physical Layer
NeedsSimple, but robust modulation, transmission, and receiving technique
Transmission mediaRadio
ISM (Industrial, Scientific, Medical) 915MHz band widely suggested
InfraredOptical media
Open research issueModulation scheme
Needed simple and low-power modulation scheme
Hardware designTiny, low-power, low-costPower efficient hardware management strategy
17APNOMS 2005 Tutorial, Okinawa, Japan
Data Link Layer
Medium Access Control (MAC)Creation of the network infrastructureFair and efficient sharing of communication resources between sensor nodes
Reasons: Existing MAC Protocols Cannot be UsedCellular system
Mobile node is only a single hop away from nearest base stationGoal : high QoS, bandwidth efficiencyImpractical for sensor network due to no central controlling agentPower efficiency is of prime importance in sensor network
Bluetooth and MANETClosest peers to sensor networksSensor network may have a much larger number of nodes
Transmission power Transmission range
18APNOMS 2005 Tutorial, Okinawa, Japan
Data Link Layer (cont.)
MAC for Sensor NetworkSMACS (Self-Organizing Medium Access Control for Sensor Networks) and EAR (Eaves-drop-And-Register) AlgorithmCSMA-Based MAC schemeHybrid TDMA/FDMA-Based MAC scheme
Power Saving Modes of OperationError Control
FEC (Forward Error Correction)ARQ (Automatic Repeat Request)
Open Research IssueMAC for mobile sensor networksPower-saving modes of operationError control coding schemes
19APNOMS 2005 Tutorial, Okinawa, Japan
Data Link Layer (cont.)
Contention-based random access
Centralized freq. and time division
Fixed allocation of duplex time slots at fixed freq.
Method
Uses constant “listening time”to min energy
Min energy use in Hardware. Picks the mix of TDMA/FDMA to min energy use
Random wake up during setup and turning radio off while idle. Exploits large bandwidth available compared to sensor data rate
Power Conservation
Should use more computation over handshaking. Use more Power-saving modes (sleeping). Need to derive bounds on energy needed by sensorsCSMA-based
Hybrid TDMA/FDMA
Not great for very mobile networks (good for mostly stationary nodes)
SMACS and EAR
What we still need to improveMAC Protocol
20APNOMS 2005 Tutorial, Okinawa, Japan
21APNOMS 2005 Tutorial, Okinawa, Japan
Network Layer
NeedsData Routing
RequirementPower efficiency Data-centric : What is the temperature on Region A ?Data aggregation
Region A
SinkSink
Query is sent to no t node but region
< Data Centric >
22APNOMS 2005 Tutorial, Okinawa, Japan
Network Layer (Cont.)
Energy-efficient route
B (PA=2)
F (PA=4)
Sink
A (PA=2)
C (PA=2)
D (PA=3)
E (PA=1)
α1 = 1
α2 = 1
α3 = 2
α4 = 2
α5 = 2
α6 = 2
α7 = 1α8 = 2
α9 = 2α10 = 2
T
Route 1 : Sink-A-B-T, total PA = 4, total α = 3Route 2 : Sink-A-B-C-T, total PA = 6, total α = 6Route 3 : Sink-D-T, total PA = 3, total α = 4Route 4 : Sink-E-F-T, total PA = 5, total α = 6
Maximum PA route : Route 4Minimum energy(ME) : Route 1Minimum hop(MH) : Route 3
23APNOMS 2005 Tutorial, Okinawa, Japan
Sensor Network Layer Routing Scheme
Sets up gradients for data to flow from source to sink during interest dissemination
Directed diffusion
Forms clusters to minimize energy dissipationLEACH
Creates multiple trees where the root of each tree is one hop neighbor from the sink
SAR
Sends data to sensor nodes only if they interested;
- has three types of messages (i.e., ADV, REQ, and DATA)SPIN
Sends data to one randomly selected neighborGossiping
Broadcasts data to all neighbor nodes regardless if they receive it before or not
Flooding
DescriptionNetwork layer scheme
24APNOMS 2005 Tutorial, Okinawa, Japan
SPIN
Step1
ADV
Step3
DATA
Step2
REQ
Step4
ADV
Step5
REQ
Step6
DATA
SPIN (Sensor Protocols for Information via Negotiation)
25APNOMS 2005 Tutorial, Okinawa, Japan
Directed Diffusion
SinkSource
Step 1 : propagate interest
SinkSource
Step 2 : set up gradient
SinkSource
Step 3 : send data
26APNOMS 2005 Tutorial, Okinawa, Japan
Transport Layer
NeedsMaintain the flow of data, and end-to-end connection if the sensor networks applications requires it
ResearchTCP splitting needed
TCP connections are ended at sink nodes, and special transport layer protocol can handle the communications between the sink node and sensor nodesCommunication between user and sink node→ TCP or UDP via the Internet or satellite
Communication between sink node and sensor node→ UDP type protocol, because sensor node has limited memory
Not based on global addressingAttribute-based naming
Open Research IssueUDP type protocol used for communication between sink node and sensor nodePower consumption, scalability and data-centric routing may need different transport handling in sensor networks
27APNOMS 2005 Tutorial, Okinawa, Japan
Application Layer
Needs Depending on the sensing tasks, different types of application software built and used
Application layer protocolsSMP (Sensor Management Protocol)
System administrators interact with sensor networks using SMPProvides the software operations needed to perform the followingadministrative tasks
Introducing the rules related to data aggregation, attribute-based naming, and clustering to the sensor nodesExchanging data related to the location finding algorithmsTime synchronization of the sensor nodesMoving sensor nodes Turning sensor nodes on / off
TADAP (Task Assignment and Data Advertisement Protocol)SQDDP (Sensor Query and Data Dissemination Protocol)
Standardization Activity
IEEE 1451IEEE 802.15.4ZigBee6LoWPAN
29APNOMS 2005 Tutorial, Okinawa, Japan
IEEE 1451 Wireless Smart Transducer Interface Standard
A new family of standards for connecting smart transducers to networks
A single communication protocol usable by all sensorsIEEE 1451.1 Network Capable Application Processor (NCAP) Information model for smart transducersIEEE 1451.2 Transducer to Microprocessor Communication Protocols and Transducer Electronic Data Sheet (TEDS) formatsIEEE1451.3 Digital Communication and Transducer Electronics Data Sheet (TEDS) Formats for Distributed Multidrop SystemsIEEE 1451.4 Mixed-mode Communication Protocols and Transducer Electronic Data Sheet (TEDS) FormatsIEEE Instrumentation and Measurement Society
30APNOMS 2005 Tutorial, Okinawa, Japan
IEEE 802.15.4 Low-Rate WPAN Standard
Key FeatureDefine the PHY and MAC spec. for low data rate wireless connectivity with fixed, portable and moving devices with no battery or very limited battery consumptionData rates: Maximum 250kbps, and minimum 20kbpsSupport for low latency devices, such as wireless keyboards, mice and joysticks, and high latency applications such as environmental sending MAC : A beacon mode for low delay with guaranteed bandwidth, and a non-beacon CSMA/CA mode for direct, mutual communicationSupports both 64-bit and 16-bit addressesSupport Multi-level securityAutomatic network establishment by the coordinatorFully handshaked protocol for transfer reliabilityPower management to ensure low power consumption
3 bands, 27 channels868MHz : 1 channel, Europe, 915MHz : 10 channels, USA2.4GHz : 16 channels, worldwide
IssuesLong battery lifeSelectable latency for controllers, sensors, remote monitoring and portable electronics
31APNOMS 2005 Tutorial, Okinawa, Japan
ZigBee Protocol Stack
PHY LAYER
MAC LAYER
NETWORK/SECURTIYLAYERS
APPLICATION FRAMEWORK
APPLICATION/PROFILES
IEEE 802.15.4
ZigBee AlliancePlatform
ZigBee or Manufacturer(User Defined)
ZigBee takes full advantage of a powerful physical radio specified by IEEE 802.15.4ZigBee adds logical network, security and application software
ZigBee characteristicsDual physical layer (2.4GHz, 869/915 MHz)Data rates of 250 kbps (@2.4 GHz), 40 kbps (@ 915 MHz), and 20 kbps (@868 MHz) Low power (battery life multi-month to years) Multiple topologies: star, peer-to-peer, mesh Addressing space of up to: -18,450,000,000,000,000,000
devices (64 bit IEEE address)- 65,535 networks
Fully hand-shaked protocol for transfer reliability Range: 50m typical (5-500m based on environment)
32APNOMS 2005 Tutorial, Okinawa, Japan
IETF 6LoWPAN
Provide IPv6 over Low Power Wireless Personal Area Network (LoWPAN), 2004 Nov., IETF 6LoWPAN WGStandard Documents
IPv6 over LoWPAN networks Goals and Assumptions: OverviewTransmission of IPv6 Packets over IEEE 802.15.4 WPAN Networks
How to transfer IPv6 packet using IEEE 802.15.4IPv6 address configuration method, header Compression, AODV-based Ad-Hoc Routing
IssuesIP adaptation / Packet Formats and interoperability Addressing schemes and address management Network management Routing in dynamically adaptive topologies Security, including set-up and maintenance Application programming interface Discovery (of devices, of services, etc) Implementation considerations
Location and Synchronization Technology in Sensor Network
34APNOMS 2005 Tutorial, Okinawa, Japan
Location Technology in Sensor Network
Discovery of absolute or relative locationGeographical routing (location attribute based naming and addressing)Tracking of moving objectsContext (location) aware applications
Challenges in Sensor NetworksEnergy constraintHarsh environment with multipathsMinimal infrastructure (Few beacons, No backend computation)
Many techniques for location sensingTDOA (Time Difference Of Arrival)TOA (Time Of Arrival)AOA (Angle Of Arrival)SSR (Signal Strength Ranging)GPS, etc.
D3-D2 D3-D1
D3D2
D1
Reader 3(x3,y3) (x2,y2)
Reader 2
Reader 1(x1,y1)
tag
(xM,yM)
TDOA Locating Algorithm
35APNOMS 2005 Tutorial, Okinawa, Japan
Time Synchronization in Sensor Network
Critical at many layers of sensor networkCommunication, localization, distributed DSP, etc.Conventional approaches
GPSNot suitable for indoors, cost, size, energy
NTP (Network Time Protocol)Delay and jitters due to MAC and store-and-forward relayingDiscovery of timer servers (nodes synchronize with one of a pre-specified list of time servers)
Reference-broadcast synchronization (RBS)Very high precision sync with slow radios
Beacons are transmitted, using physical-layer broadcast, to a set of receiversTime synchronization is based on the difference between reception times, do not sync sender with receivers
USN Management Requirements
37APNOMS 2005 Tutorial, Okinawa, Japan
User Requirements on Wireless Sensor Network
A survey on the characteristics of a wireless sensor network which are most most important to users (2002)1. Data Reliability2. Battery Life3. Cost 4. Transmission Range5. Date Rate6. Data Latency 7. Physical Size8. Date Security
38APNOMS 2005 Tutorial, Okinawa, Japan
User Requirements on Wireless Sensor Network (Cont.)
A survey on the characteristics of a wireless sensor network which are most most important to users (2002)1. Data Reliability Reliability, Configuration
(Density), Security Management2. Battery Life Power Management3. Cost 4. Transmission Range, 5. Date Rate,6. Data Latency
Performance Management7. Physical Size8. Date Security Security Management
39APNOMS 2005 Tutorial, Okinawa, Japan
USN Management Requirements
Fault toleranceHandle loss of nodes - Lack of Power, Physical damage, Environmental interference
ScalabilityHandle high density of nodes - The number of sensor nodes is an extreme value of millions
Production costsMake them low cost - Cost of a single node is very important to justify the overall cost of the network
Operating environmentSurvive and maintain communication - The bottom of an ocean, biologically contaminated field, battlefield
Transmission mediaWireless - Radio, infrared, optical media
Hardware constraintsNodes are tiny - Very small size, very light node, limited memory, limited battery
Power consumptionLimited Tx, computation, lifetime - Replenishment of power is impossible
Changing TopologyNodes - Nodes moving, new nodes, loss nodes
40APNOMS 2005 Tutorial, Okinawa, Japan
Ubiquitous Sensor Network Management
Why isn’t SNMP (Simple Network Management Protocol) adaptable to USN?
Sensor-specific failures are not handledDifficult to find the failed nodesPhysical connections are not utilizedCommonly, there is not a management agentSpecifying nodes is difficultNetwork is self-configured, so that management server doesn’t have all information of sensor nodes
ChallengesPresents many and drastically different challenges. For example:
Deployment of nodes, Discarding of nodes Requires augmentation to (or new approaches over) traditional network and service management techniques Needs to take into account specific characteristics of WSNs (e.g., energy waste)
USN Management Architecture & Functions
42APNOMS 2005 Tutorial, Okinawa, Japan
Sensor Network Management Functional Areas:
Reliability
Management
Reliability
Management
Context
Management
Context
Management
Topology
Management
Topology
Management
Security
Management
Security
Management
Performance
Management
Performance
Management
Power Management
Context
Management
Context
Management
43APNOMS 2005 Tutorial, Okinawa, Japan
USN Management Architecture
To
po
log
y
Security ManagementPrivacy Management
Self-Configuration,Configuration Management
Application
Transport
Network
Data Link
Physical Layer
Po
wer
Relia
bility
Secu
rityPower ON/OFF, Power ExhaustingEnergy-efficient ProtocolEssential Management Function
Context-Awareness ManagementTask Management, Location-Based Context Management
Fault / Reliability Management, Performance Management
Co
nte
xt
Related Research on USN Management Architecture
45APNOMS 2005 Tutorial, Okinawa, Japan
Ad-Hoc Network Management Characteristics
Dynamic Topology ChangesPower Limitation
Power management strategy affects topological configuration, and fault management.The faulty conditions are different from those of wire-line network or wireless networks.
Constrained Wireless CommunicationLarge Number of Heterogeneous Nodes
46APNOMS 2005 Tutorial, Okinawa, Japan
ANMP (Ad-Hoc Network Management Protocol)
ANMPA protocol for managing mobile wireless ad-hoc networks
Focuses on data collection, configuration, fault and security management
Uses hierarchical clustering of nodes Helps to reduce exchanges between manager and agents Easier to keep track of roaming nodes
Fully compatible with SNMPv3 Includes enhanced security features
ArchitectureHierarchical3-Level
Manager
Cluster head
Clusters
Agents
47APNOMS 2005 Tutorial, Okinawa, Japan
ANMP: Data Collection
ANMP ClusteringFormed for management purposes only - different from those formed for routing purposes Dynamic structures – number and composition of nodes change over time Nodes acting as cluster head may also change Two algorithms proposed for clustering:
Graph based clustering (graphic view of the net; each node is nomore than one or two hops away from the cluster head) Geographical clustering (based on spatial density of nodes usinglatitudes and longitude information)
Data CollectionEvery node runs ANMP locally A new MIB, called ANMP MIB added to MIB
anmpMIB
Powerusage(1)
topologyMaintenance(2)
agentsInformation(3)
Lacm(4)
ANMP MIB
48APNOMS 2005 Tutorial, Okinawa, Japan
Guerrilla Management
Guerrilla ManagementA supervisor/agency architecture for scalable and cooperative management
Uses mobile code techniques for nomadic and active management
Uses a dynamic adaptive protocol for clustering and selecting nomadic managers
Nodes range in functionality and capability SNMP-capable, Probe-capable, and Full-featured
49APNOMS 2005 Tutorial, Okinawa, Japan
Guerrilla Management Model
SupervisorSupervisor
Nomadic ManagerNomadic Manager
Nomadic ManagerNomadic Manager Nomadic ManagerNomadic Manager
Nomadic ManagerNomadic Manager
Agency
ActiveProbesActiveProbesActive
ProbesActiveProbes
ActiveProbesActiveProbes
ActiveProbesActiveProbesActive
ProbesActiveProbes
ActiveProbesActiveProbes
ActiveProbesActiveProbesActive
ProbesActiveProbes
ActiveProbesActiveProbes
High-Tier
Low-Tier
50APNOMS 2005 Tutorial, Okinawa, Japan
Classification of Nodes in GuerrilaManagement Architecture
MIBSNMP agent
Communication protocolCommunication protocol
MIBSNMP agent
Communication protocolCommunication protocol
Probe processingmodule
Probe processingmodule
MIBSNMP agent
Communication protocolCommunication protocol
Probe processingmodule
Probe processingmodule
GMIBNormadic management module
a) SNMP-capable node b) probe-capable node architecture
c) full-featured node architecture
51APNOMS 2005 Tutorial, Okinawa, Japan
GMIB: Guerrilla MIB
Nomadic ManagerCollaborates autonomously to manage the entire ad hoc network with minimal help from the supervisor
Role change
Load sharing - cloning itself into another node
Spawning and merging
GMIB (Guerrilla MIB) A data structure equivalent to a SNMP MIB
An aggregation of management information collected from neighbor nodes via probes
Maintained inside NMM
Also includes Management information (e.g., neighbor information) in the probeprocessing modules
Can be accessed by both the NMM and incoming probes Modeled as a branch in a SNMP MIB
52APNOMS 2005 Tutorial, Okinawa, Japan
MANNA
Wireless Sensor Network (WSN) Functionalities Another abstraction level to include the network functionalitiesUseful in developing various network management models
Identifies a set of WSN-specific Managed Objects mostly derived from OSI
Management ArchitecturesFunctionalInformationPhysical
53APNOMS 2005 Tutorial, Okinawa, Japan
MANNA Architecture
Functional ArchitectureDescribes the distribution of management functionalities among manager, agent, and management information base (MIB) Covers variety of manager-agent models
MANNA Manager MANNA Agents
Informational ArchitectureTwo Object Class Type
Managed Object Classes Support object classes
54APNOMS 2005 Tutorial, Okinawa, Japan
Management Functionality Abstraction in Manna
ManagementFunction
FaultConfigurationPerformance
SecurityAccounting
Managementfunctionalareas
Business managementService managementNetwork managementNetwork element managementNetwork element
Managementlevels
Conf
igura
tion
Mainte
nanc
eSe
nsing
Proc
essin
gCo
mmunica
tion
WSN functionalities
55APNOMS 2005 Tutorial, Okinawa, Japan
USN Management Architecture (MANNA)
• Requirements that characterize a sensor network come from the objectives defined in the business management layer
• Design of WSNs depends on applications
• The larger the number of monitored parameters, the larger the energy consumption and the lower the network lifetime
• Basic USN services are Sensing, Processing, and Dissemination
• Aims to manage a network as a whole, which is typically distributed over an extensive geographical area
• The relationships among sensor nodes are to be considered: collaboration, connectivity, and correlation
• Power management• Mobility management• State management
• Physical resource: power supply, processor, memory, sensor device, and transceiver
• Logical resource: communication protocols, application programs, correlation procedures, operating systems, and network services
Business Management
Service Management
Network Management
Element Management
Network Entities
56APNOMS 2005 Tutorial, Okinawa, Japan
USN Management Functional Area (MANNA)
Security functionalities for USNs are intrinsically difficult to be provided because of their ad-hoc organization, intermittent connectivity, wireless communication and resource limitationsA WSN is subject to different safety threats: internal, external, accidental, and malicious
Security
There is a trade-off to be considered: the higher the number of managed parameters, the higher the energy consumption and the lower the network lifetimeOn the other hand, if enough parameter values are not obtained, it may not be possible to manage the network appropriately
Performance
It includes functions related to the use of resources and corresponding reportsIt establishes metrics, quotas and limits that can be used by functions of other functional areasIt must provide self-sustaining functionalities
Accounting
Self-organization: is the property which the sensor nodes must have to organize themselves to form the networkSelf-configuration: nodes setup and network boot up must occur automatically
Configuration
Faults in USNs are not an exception and tend to occur frequently, thus fault management is a critical function.This is one of the reasons that make WSN management different from traditional network managementSelf-diagnostic: the network monitors itself and find faulty or unavailable nodesSelf-healing: the network prevents disruptions or that acts to recover itself or the node after the self-diagnostic
Fault
Functions
Specific Management Functions of USN
Power ManagementTopology ManagementSecurity ManagementContext Management
58APNOMS 2005 Tutorial, Okinawa, Japan
Power Management
Manages how a sensor node uses its power
ExampleSensor node may turn off its receiver after receiving a message from one of its neighbors
Avoid getting duplicated messages
When the power level of the sensor node is lowBroadcasts to its neighbor when it is low in power
Cannot participate in routing messages
Reserve the remaining power for sensing
RequirementsUsing batteryLimited PowerExpand the life time of sensor nodeReduce the overhead
59APNOMS 2005 Tutorial, Okinawa, Japan
Power Management in Protocol Layer
Physical layerLow Power Modulation SchemeTransceiver, Sensor, Processor: Small, Low Power, Low Cost
Data link layerEnergy efficiency MAC protocol
Adaptive duty cycling – SMAC, ASCENT, SPANWake up on-demand – STEM, Wake-on-Wireless
Reduce the collision, signaling, frame overheadPower saving mode (ex. On/Off mode)
Network LayerEnergy-efficiency routingEnergy-efficiency data aggregation algorithmsLocation-based routing
Transport LayerUse UDP message protocol between Sink and Sensor nodeLimited memory and processing power
Application LayerEnergy-efficiency applications
60APNOMS 2005 Tutorial, Okinawa, Japan
Topology Management
GoalCoordinate the sleep transitions of all nodes, while ensuring adequate network connectivity, such that data can be forwarded efficiently to the data sink
RequirementsHeterogeneous nodeData discovery & data disseminationLimited memory & power constraintApplication requirementsNode mobility
Ad-hoc Self-organization LCA (Linked Cluster Algorithm)LAA (Link Activation Algorithm)DEA (Distributed Evolution Algorithm)
61APNOMS 2005 Tutorial, Okinawa, Japan
Topology Management (Cont.)
SMACS (Self-Organizing Medium Access Control for Sensor networks)EAR (Eavesdrop And Register)
BI (Broadcast Invite)MI (Mobile Invite)MR (Mobile Response)MD (Mobile Disconnect)
SAR (Sequential Assignment Routing)SWE (Single Winner Election)MWE (Multi Winner Election)
62APNOMS 2005 Tutorial, Okinawa, Japan
Security Management
RequirementsPeanut CPU (slow computation rate)Battery power: trade-off between security and battery lifeLimited memoryHigh latency: conserve power, turn on periodically
Security Management in USNApplications need security (privacy)Absence of security enables attacks such as spoofing & replay attacks, resulting in DoS or system compromiseIntrusion prevention : First line of defenseIntrusion detection : Second line of defense
Main Security Threats in USNRadio links are insecureSensor nodes are not tamper resistant – it it is compromised, attacker obtains all security information
Attacker typesMote-class: attacker has access to some number of nodes w/ similar characteristicsOutside / inside
63APNOMS 2005 Tutorial, Okinawa, Japan
Security Management (Cont.)
AttacksPhysical attackDenial-of-serviceBattery exhaustionClock synchronizationLocation discoveryAttacks on routing
spoofed, altered, or replayed routing informationselective forwarding by compromised nodessinkhole attack: pretend to be a sink node (fake sink node)sybil attack: a single node pretending to be in different parts of the networkWormholes: tunnel packets received on one part of the network toanotherHELLO flood attacksacknowledgment spoofing
64APNOMS 2005 Tutorial, Okinawa, Japan
CountermeasuresLink layer encryption – selective forwardingUsing a counter – Replay attacksLimiting the number of neighbors per node – Inside attacksBi-directionality of the link – HELLO floodGeographical routing – Wormhole attacks
Security Management (Cont.)
65APNOMS 2005 Tutorial, Okinawa, Japan
Context Management
Gathering the “User Context”Requirement
User intent predictionApplication deployment supportRuntime context serviceReal-time serviceInter-user coordination and collaboration
ContextAny information that can be used to characterize the situation of an entity
Considered relevant to the interaction of an entityConsidered relevant to the interaction between a user and an application, including themselves
Context ModelACTIVITY – behavior, taskENVIRONMENT
physical status (location, time, etc)social surroundings (nearby objects)
SELF – status of device itself
Activity
Environment Self
66APNOMS 2005 Tutorial, Okinawa, Japan
Context Management (cont’)Key Components
Context discovery and acquisitionUser interfaceContext management and modelingContext composition and gathering
Group Context ManagementEnable syntactic and semantic interoperability between context- aware applicationsEnable seamless integration of various kinds of contexts and make it easy to be inferred
User ContextUser intent predictionApplication development supportRuntime context serviceInter-user coordination and collaboration
67APNOMS 2005 Tutorial, Okinawa, Japan
ConclusionBrief Overview on Ubiquitous Sensor Network Technology
Identify salient features of USN Introduce sensor network protocols Current standardization activities
Analyze USN Management RequirementsPropose a USN Management ArchitectureIntroduce USN Management Functions
Related Research on USN Management ArchitectureIdentify Specific USN Management Functions
Power ManagementTopology ManagementSecurity ManagementContext Management
Research of sensor network management are at the early stages and a challenging task
68APNOMS 2005 Tutorial, Okinawa, Japan
References
Sensor Network Projects UC Berkeley: Smart Dust and TinyOS: today.cs.berkeley.edu/tos; robotics.eecs.berkeley.edu/~pister/SmartDust; also www.tinyos.netEyes: http://eyes.eu.org/index.htmCornell: Cougar: http://www.cs.cornell.edu/database/cougar/UCLA: CENS - Center for Embedded Networked Sensing www.cens.ucla.edu/ Northern Arizona: WNRL: www.cet.nau.edu/Projects/WNRL USC: SCADDS: www.isi.edu/scaddsUCLA: WINS: www.janet.ucla.edu/WINS JPL: Sensor Web: sensorwebs.jpl.nasa.gov/ Georgia Tech: SensorSimII: users.ece.gatech.edu/~grimace/research/sensorsimii/index.html
Basics of Sensor NetworkIan F. Akyildiz, W, S. Weilian, Y. Sankarasubramaniam, and E. Cayirci, “A Survey on Sensor Networks”, Communication Magazine IEEE. vol. 40, no. 8, pp. 102-114, Aug. 2002.A. WADAA, S. OLARIU and L. WILSON, M. ELTOWEISSY, K. JONES “Training a Wireless Sensor Network”Macros Augusto M. Vieira, Claudionor N. Coelho. Jr., Diogenes Cecilio da Silva Junior, Jose M. da Mata, “Survey on Wireless Sensor Network Devices”Chien-Chung Shen, Chavalit Srisathapornphat, Chaiporn Jaikaeo, “Sensor Information Networking Architecture and Applications” IEEE Personal Comm., August, 2001M. Ulema, Wireless Sensor Networks; Applications, Technology, and Management, IEEE NOMS 2004, Tutorial
69APNOMS 2005 Tutorial, Okinawa, Japan
References (Cont.)Ubiquitous Sensor Network Management
Wang Feng, Tian Qichuan, Gao Quanzue and Pan Quan, “A Study of Sensor Management Based on Sensor Networks”, International Conference on Robotics, Intelligent Systems and Signal Processing, October, 2003Mark A. Perillo, Wendi B. Heinzelman, “Optimal Sensor Management Under Energy and Reliability Constraints,” IEEE, 2003W. Chen, N. Jain, and S. Singh, “Anmp: Ad Hoc Net-work Network Management Protocol,”IEEE JSAC, vol. 17,no. 8, Aug. 1999. C-C. Shen, C. Srisathapornphat, and C. Jaikaeo, “An Adaptive Management Architecture for Ad hoc Networks,” IEEE Communications Magazine, Feb. 2003. I.F.Akyildiz, W.Su , Y.Sankarasubramaniam, E.Cayirci, “Wireless sensor networks: survey,”Computer Networks 38 (2002)393 –422 D. Estrin, “Some Distributed Coordination Schemes for Wireless Sensor Networks,” Talk given at StanfordNov 2000 L. B. Ruiz, J. M. S. Nogueira, and A. A.F. Loureiro, “MANNA: Management Architecture for Wireless Sensor Networks,” , IEEE Communications Magazine, Feb. 2003.
Topology Management Mirkovic, J.; Venkataramani, G.P.; Lu, S., Zhang, L., “A self-organizing approach to data forwarding in large-scale sensor networks,“ ICC 2001. IEEE International Conference, June 2001Alberto Cerpa and Deborah Estrin, “ASCENT: Adaptive Self-Configuring Sensor Networks Topologies,” in Proceedings of the Twenty First International Annual Joint Conference of the IEEE Computer and Communications Societies (INFOCOM 2002), June, 2002Katayoun Sohrabi, Jay Gao, Vishal Ailawadhi, and Gregory J. Pottie, “Protocols for Self-Organization of Wireless Sensor Network,” IEEE Personal Communications, October, 2000
70APNOMS 2005 Tutorial, Okinawa, Japan
References (Cont.)Context-Awareness Management
“A Middleware Infrastructure for Active Surroundings”, Pervasive Computing Group, Technical Report No. CSPG-2003 04 028James Powell, “Middleware for Pervasive and Proactive Computing”, 2003William Plymale, James Powell, “Pervasive Computing and Institutional Repositories”
Security Management A. Perrig, R. Szewczyk, VictorWen, D. Culler, and J. D. Tygar. “SPINS: Security protocols for sensor networks.” In Proceedings of Seventh Annual International Conference on Mobile Computing and Networks MobiCom 2001, July 2001.J. Douceur. “The sybil attack.” In Proceedings of the IPTPS 2002, Cambridge, MA, USA, March 2002.Yih-Chun Hu, A. Perrig, and D. B. Johnson. “Wormhole detection in wireless ad hoc networks. Technical report,” Department of Computer Science, Rice University, December 2001.Technical Report TR01–384C. Karlof and D. Wagner, “Summary of “Secure Routing in Wireless Sensor Networks: Attacks and Countermeasures”
CompanyEmber: www.ember.comCrossbow: www.xbow.comMillenial Net: www.millennial.netDust Inc. : www. dust-inc.comLuna iMonitoring: www.lunaimonitoring.comMicroStrain: www.microstrain.comSensoria Corp.: www.sensoria.comXsilogy: www.xsilogy.comZigBee Alliance: www.zigbee.org
71APNOMS 2005 Tutorial, Okinawa, Japan
References (Cont.)
Other ReferencesA. Bharathidasan and V. Ponduru, “Sensor Networks: An Overview” , Department of Computer Science, University of California, Davis, CA 95616 TinyOS web site: http://www.tinyos.net/ Tim Nieberg , ”Wireless Sensor Networks: The EYES Project,” Ubiquitous Computing Colloquia 2003, Universiteit Twente, Apr. 11, 2003 H. Karl, “Making sensor networks useful: Distributed services,” ESF Workshop Oct. 02 2003. Sohrabi, K, Gao, J., Ailawadhi, V, and Pottie, G., "Protocols for self-organization of a wireless sensor network," IEEE Personal Comm. Magazine, vol. 7, No. 5, pp. 16-27, Oct. 2000L. St. Ville, “An architectural overview of a Distributed Sensor Network with Mobile Sensors,”February 2001A. Mainwaring, J. Polastre, R. Szewczyk, and D. Culler, “Wireless Sensor Networks for Habitat Monitoring,” ACM International Workshop on Wireless Sensor Networks and Applications, 2002. W. R. Heinzelman, J. Kulik, and H. Balakrishnan, “Adaptive Protocols for Information Dissemination in Wireless Sensor Networks,” Proc. ACM MobiCom ’99, Seattle, WA, 1999 S. Tilak, N. Abu-Ghazaleh, and W. Heinzelman, "A Taxonomy of Wireless Micro-Sensor Network Models", ACM Mobile Computing and Communications Review (MC2R), Volume 6, Number 2, April 2002 D. Estrin, “Sensor Network Protocols Tutorial,’ Mobicom 2002A. Woo, and D. Culler, “A Transmission Control Scheme for Media Access in Sensor Networks,”Proc. ACM MobiCom ’01, Rome, Italy, July 2001, Craig Ulmer, “Wireless Sensor Networks,” Presentation at NASA's Jet Propulsion Lab, August 10, 2000 Mani Srivastava, “Sensor Node Platforms & Energy Issues,” Tutorial, Mobicom 2002 Wei Ye, John Heidemann, Deborah Estrin, ”An Energy-Efficient MAC Protocol for Wireless Sensor Networks”, Infocom 2002
72APNOMS 2005 Tutorial, Okinawa, Japan
Questions ?
73APNOMS 2005 Tutorial, Okinawa, Japan
Thank you!