1 / 35

Transpo rt Protocols in

Transpo rt Protocols in. Wireless Sensor Networks. Kevin Mendes Lakehead University. Topics to be Covered. Trans p ort Protocol O bjectives Draw b acks of TCP in W SNs Design Guid e li n es ( Performa n ce Metrics ) Cong e stion Control & its Causes Loss Recove r y.

senona
Download Presentation

Transpo rt Protocols in

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. TransportProtocolsin WirelessSensorNetworks Kevin Mendes Lakehead University

  2. Topics to be Covered • TransportProtocol Objectives • Drawbacksof TCPin WSNs • DesignGuidelines (PerformanceMetrics) • CongestionControl & its Causes • LossRecovery

  3. TransportProtocol Objectives • Transportprotocolsare usedto: • Mitigatecongestion, • Reducepacketloss, • Providefairnessin bandwidthallocation, • Guaranteeend-to-endreliability. • Thetraditionaltransport protocols (i.e.,UDPandTCP) cannot be directlyimplementedfor WSNs • UDPdoesnotprovidedeliveryreliabilitythatisoftenneeded formanysensorapplications, • Nordoesitofferflow andcongestioncontrolthatcanlead to packetlossandunnecessaryenergyconsumption.

  4. Drawbacksof TCPin WSNs • Overheadassociatedwith TCP connection • establishment • Flow andcongestioncontrol mechanismsinTCP • Resultin unfairbandwidthallocationanddatacollections. • TCPassumesthatpacketlossisduetocongestion • TCPhasadegradedthroughputin wirelesssystems • TCPrelieson end-to-endretransmission • Consumesmoreenergyandbandwidththanhop-by-hop • retransmission. • TCPguaranteessuccessfultransmissionof packets • Isnotalwaysnecessaryforevent-drivenapplicationsin • sensornetworks.

  5. DesignGuidelines • In WSNsseveralnewfactors,canresultin congestion: • Convergentnature ofupstreamtraffic • Limited wirelessbandwidth • TworeasonsofpacketlossinWSNs: • Packet loss dueto congestioninintermediate nodes • Packet loss dueto bit-error rate(number of bit errors divided by the total number of transferred bits during a studied time interval)inwirelesschannel • TwomajorproblemsthatWSNtransport protocolsneedtocopewith: • Congestion • Packet loss.

  6. PerformanceMetrics • TransportprotocolsforWSNsshouldprovide: • End-to-endreliability • End-to-endQoS • Performancemetrics: • Energy efficiency, • Reliability, • QoS • Packet-lossratio, • Packet-delivery • Latency • Fairness.

  7. PerformanceMetrics: EnergyEfficiency • Sensornodeshave limitedenergy. • Transportprotocolsshouldmaintainhighenergy efficiency • Tomaximizesystemlifetime. • Forloss-sensitiveapplications, • Packetlossleadstoretransmission • Inevitableconsumptionofadditional batterypower • Therefore,severalfactorsneedtobecarefully • considered, • Numberofpacketretransmissions, • Distance(e.g.,hop)foreachretransmission, • Overheadassociatedwithcontrolmessages.

  8. PerformanceMetrics: Reliability • ReliabilityinWSNscanbeclassifiedintothe • following categories: • Packet reliability: • Applicationsareloss-sensitiveandrequiresuccessful transmissionofall packetsorata certainsuccessratio. • Event reliability: • Applicationsrequireonly successfuleventdetection,but • notsuccessfultransmissionofallpackets. • Destination-relatedreliability: • Messagesmightneedtobe deliveredtosensornodes: • Thatarein aspecificsubarea • Thatareequippedwithaparticularsensortype.

  9. PerformanceMetrics: QoSMetrics • QoSmetricsinclude: • Bandwidth, • Latencyor delay, • Packet-loss ratio. • Dependingontheapplication,thesemetricsor theirvariantscouldbeused forWSNs. • TargettrackingApplication: • Generatehigh-speeddatastreams • Requirehigherbandwidth • Fora delay-sensitiveapplication: • Mayalsorequiretimelydeliverydata.

  10. PerformanceMetrics: Fairness • What if Sensornodesarescatteredinageographical • Area? • Many-to-oneconvergentnatureofupstream • traffic: • Itis difficult forsensornodesthatare farawayfrom thesinkto transmitdata. • Transportprotocolsneedtoallocate • bandwidthfairlyamongallsensornodes • Sink can obtaina fairamount of data fromall the sensor nodes.

  11. CongestionControl  Closed-loopcontrol feedback Feedbackshouldbefrequent,butnottoomuchotherwisetherewillbeoscillations Cannotcontrolthebehaviorwith atimegranularitylessthanthefeedbackperiod

  12. EffectofCongestion • Packetloss • Retransmission • Reducedthroughput • Congestioncollapsedue to • Unnecessarilyretransmittedpackets • Undeliveredorunusablepackets

  13. CongestionControlintheInternet improving the efficiency of TCP/IP networks by reducing the number of packets that need to be sent over the network. ActiveQueue Management(AQM) TCPCongestionControl

  14. CausesFor Congestion in WSNs • Duetothepacket-arrivalrateexceeding • thepacket-servicerate. • Thisismorelikelytooccuratsensornodes • closeto thesink • Link-levelperformanceaspectssuchas: • Contention, • Interference, • Bit-errorrate. • Thistypeof congestionoccursonthelink.

  15. Typesof Congestionin WSNs • NodeLevel Congestion: • It isduetothepacket-arrival rateexceedingthepacket- servicerate. • Thisis morelikelytooccur at sensornodesclosetothe sink. • LinkLevel Congestion: • It aspects suchas contention,interference,and bit-errorrate.

  16. EffectsofCongestion in WSNs • Energy: • Wastethelimitednodeenergy • ApplicationQoS: • Degrade reliabilityandapplicationQoS • Buffer overflow • Larger queuingdelays • Higherpacket loss. • Degrade linkutilization. • ItresultsintransmissioncollisionsifCSMA,is • used • increases packet-servicetime • wastes energy.

  17. CongestionControlApproaches • Therearetwogeneralapproachesto • controlcongestion: • Networkresourcemanagement: • triesto increasenetwork resource to mitigate congestion • Trafficcontrol: • impliesto control congestionthroughadjusting trafficrate atsource nodesor intermediates nodes

  18. TrafficControlMethods • End-to-end: • Can imposeexact rate adjustment at each source node • Simplify the designat intermediatenodes • Itresults inslowresponseandrelieshighlyonthe • round-triptime (RTT). • Hop-by-hop: • Ithas fasterresponse. • Difficultto adjust thepacket forwardingrate at intermediate nodes • BecausepacketforwardingrateisdependentonMAC • protocolandcouldbe variable.

  19. CongestionControlParts • Congestiondetection • Monitor buffer/queue size • Monitor channelbusy time, estimatechannel’sload • Monitor theinter-packet arrivaltime (data,ctrl) • Congestionnotification • Explicitcongestionnotificationinpacketheader,then broadcast (but then energy-consuming!) • RateAdjustment • Dynamic reportingrate dependingon congestionlevel • In-networkdatareductiontechniques(agressive • aggregation)oncongestion

  20. CongestionDetection • In TCP: • Congestionis observedat theendnodesbasedon atimeoutor redundantAcknowledgments. • In WSNs: • Proactivemethods are preferred. • Congestion indicators: • Queuelength • Packet service time • Theratioof packet servicetimeover packet • interarrivaltime

  21. CongestionNotification • Propagationofcongestioninformationfromthe • congested node • Totheupstreamsensornodes • Tothesourcenodesthatcontributetocongestion • Congestioninformation • CongestionNotification(CN)bit, • Ormoreinformationsuchasallowable datarate,orthe congestiondegree. • Disseminatingcongestioninformation: • Explicit • Usesspecialcontrol messagesto notifytheinvolvedsensor nodes of congestion • Implicit • Piggybackscongestioninformationin normaldata packets.

  22. RateAdjustment • Uponreceivingacongestion indication,a • sensornodecanadjustitstransmissionrate. • If a singleCNbitisused: • AdditiveIncreaseMultiplicativeDecrease(AIMD) • If additionalcongestion informationis available: • Accuraterate adjustmentcanbe implemented

  23. LossRecovery • Reasonsofpacket loss in wirelessenvironments: • Congestion • Biterror • nodefailure, • wrongoroutdatedroutinginformation, • Energydepletion. • How to overcomethisproblem: • Increasethesourcesendingrate • Workswellforguaranteeingevent reliability • Isnotenergyefficient • Introduceretransmission-basedlossrecovery. • Ismoreactiveandenergyefficient • Can be implementedat both thelinkand transportlayers. • Link-layerloss recoveryis hop-by-hop,whilethetransportlayer recoveryis usuallydoneend-to-end.

  24. LossDetection andNotification • Acommonmechanismis toincludea sequencenumberineachpacket header. • Thecontinuityofsequencenumbers can beusedtodetectpacketloss. • Lossdetectionandnotificationcanbe: • End-to-end • Hop-by-hop.

  25. End-to-EndApproaches • End-points(destinationorsource)are • responsibleforlossdetectionandnotification. • Drawbacks • Isnotenergy efficient. • Thecontrolmessageswould utilize areturnpath consistingofseveralhops • Control messagestravelthroughmultiplehops • Couldbe lostwitha highprobabilityduetoeitherlinkerror orcongestion. • Leads toend-to-endretransmissionsfor loss • recovery.

  26. Hop-by-hopLossDetectionand Notification • Intermediatenodesdetectandnotifypacket • loss. • Apairofneighboringnodesareresponsible • for lossdetection. • Is moreenergyefficient. • Twocategories: • Receiver-based • Receiverinferspacketlosswhenitobservesout-of- sequencepacketarrivals. • Senderbased • Senderdetectspacketlossoneitheratimer-basedor overhearingmechanism.

  27. MethodstoNotify theSender • Specialcontrolmessages: • ACK(Acknowledgment) • NACK(NegativeACK) • PiggybackingACKinthepacket header • IACK(ImplicitACK) usingoverhearing • Avoids control messageoverhead • Moreenergyefficient. • Sensor nodesmusthavethecapabilitytooverhearthephysical • channel. • Isnotfeasiblewhen: • Transmissionis corrupt • Channelisnotbidirectional • Sensornodesaccess thephysicalchannelusingTimeDivision MultipleAccess(TDMA)-basedprotocols

  28. Retransmission-BasedLoss Recovery • End-to-end • The source performsretransmission. • Hop-byhop. • Anintermediatenodethat interceptsloss notification searchesitslocalbuffer. • Ifitfinds a copy of thelostpacket, it retransmits the packet. • Otherwiseitrelayslossinformation upstreamto • otherintermediatenodes.

  29. Comparisons • End-to-end retransmission: • Thecachepointis thesourcenode. • Hasalongerretransmissiondistance • Allowsforapplication-dependentvariablereliabilitylevels • Hop-by-hopretransmission: • Thecachepointcouldbethepredecessornodeoftheloss point. • Ismoreenergy-efficient • Requiresintermediatenodestocachepackets. • Ispreferredif100percentpacketreliabilityis required • Cannotassuremessagedeliveryin thepresenceofnode • failure

  30. IssuesRelated toHop-by-hop Retransmission • Immediateretransmission • Retransmissioncanbe triggeredimmediatelyuponthe detectionofa packetloss. • Resultsin shorterdelay • Ifpacketlossiscausedbycongestionitcouldaggravatethe • congestionsituationandcausemorepacketlosses. • Distributed TCP Cache(DTC) • Giventhelimited memoryin sensornodes,packetsmayonly • needtobecachedatselectednodes. • Howtodistributecachedpacketsamongasetofnodes? • Itbalancethebufferconstraintsandretransmissionefficiency byusingprobability-basedselectionforcachepoints.

  31. DesignGuidelines • Severalfactorsmustbetaken into consideration: • Topology • Diversityofapplications • Trafficcharacteristics • Resourceconstraints • Transportprotocolscomponents • Congestioncontrol • Lossrecovery • Twoapproaches • Designseparateprotocolsoralgorithms,respectively,for congestioncontrolandlossrecovery. • Providescongestionandlosscontrolin an integratedway • Thejointuseofthesetwoprotocolsmayprovidethefull • functionalityrequiredbythetransportprotocolsforWSNs.

  32. TheExisting TransportProtocolsfor WSNS • ProtocolsforCongestionControl • CongestionDetectionandAvoidance(CODA) • Control andFairness (CCF) • PumpSlowlyFetch Quickly(PSFQ) • Priority-basedCongestionControlProtocol(PCCP) • Siphon • Adaptive Rate Control(ARC) • Trickle • ProtocolsforReliability • ReliableMulti-Segment Transport(RMST) • ReliableBurstyConvergecast(RBC) • Event-to-Sink ReliableTransport (ESRT) • GARUDA • ProtocolsforCongestionControlandReliability • Sensor TransmissionControlProtocol(STCP)

  33. WSNCongestionControlProtocols • CCF : • Packetservicetime • Implicit • Exact hop-by-hoprate adjustment • STCP: • Queuelength, • Implicitcongestionnotification, • AIMD-like end-to-endrate adjustment • Fusion: • Queuelength, • Implicitcongestionnotification, • Stop-and-starthop-by-hoprate • adjustment • CODA: • Queuelengthand channel status, • Explicitcongestionnotification, • AIMD-likeend-to-endrate • adjustment • PCCP: • Packetinterarrivaltime and packetservicetime, • Implicitcongestionnotification, • Exact hop-by-hoprate • adjustment • ARC: • Theevent if thepackets are successfully forwardedor not, • Implicitcongestionnotification, • AIMD-like hop-by-hoprate adjustment

  34. References A Survey of Transport Protocols for Wireless Sensor Networks: Chong gang Wang and KazemSohraby, University of Arkansas Link:http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=01637930

  35. Thank You • Any Questions???

More Related