IEEE MMTC R-Letter - IEEE Communications Society

IEEE COMSOC MMTC R-LetterIn summary, the paper proposes two mechanisms toprevent the attacks for collaborative spectrum sensingin cognitive radio network. The detail and completeanalysis shows that the direct punishment mechanismcan effectively prevent all attacks in both “attack-andrun” and “stay-with-attacks,” and the indirectpunishment mechanism can prevent all attacks in thelong-run as long as the attackers care about theirfuture rewards.Fen Hou is an assistant professor in the Departmentof Electrical and Computer Engineering at theUniversity of Macau since2013. Dr. Fen Hou receivedthe Ph.D. degree in electricaland computer engineeringfrom the University ofWaterloo, Canada, in 2008.Sheworkedasapostdoctoral fellow in theElectrical and ComputerEngineeringattheUniversity of Waterloo andin the Department of Information Engineering at theChinese University of Hong Kong from 2008 to 2009and from 2009 to 2011, respectively. She worked as alecture at the Macao Polytechnic Institute from 2011to 2012. Her research interests include resourceallocation and scheduling in broadband wirelessnetworks, protocol design and QoS provisioning incognitive radio network, and mechanism design inparticipatory sensor networks. Dr. Hou is therecipient of IEEE GLOBECOM Best Paper Award in2010, as well as the Distinguished Service Award inIEEEComSocMultimediaCommunicationsTechnical Committee (MMTC) in 2011.Acknowledgement:This paper is nominated by the MMTC WirelessTechnology for Multimedia Communications(WTMC) Interest Group.References:[1] S. Haykin, “Cooperative radio: Brain-empoweredwireless communications,” IEEE Journal onSelected Areas on Communications, vol. 23, no.2, pp. 201-220, 2005.[2] I. F. Akyildiz, W. Lee, M. C. Vuran, and S.Mohanty, “NeXt generation/dynamic spectrumaccess/cognitive radio wireless networks: asurvey,” Computer Networks, vol. 50, no. 13, pp.2127-2159, 2006.[3] K. B. Letaief and W. Zhang, “Cooperativecommunications for cognitive radio networks,”Proc. IEEE, vol. 97, no. 5, pp. 878-893, 2009.[4] H. Li and Z. Han, “Catch me if you can: anabnormiality detection approach for collaborativespectrum sensing in cognitive radio networks,”IEEE Trans. Wireless Commun., vol. 9, no. 11,pp. 1554-3565, 2010.http://committees.comsoc.org/mmcDr. Hou has served as the chair of an Interest Groupin IEEE MMTC, as well as a technical programcommittee member for IEEE WiOpt 2012, IEEE ICC2011, IEEE WCNC 2011, IEEE GLOBECOM 2010,etc. Dr. Hou has also served as a technical reviewerfor IEEE Transactions on Wireless Communications,IEEE Transactions on Vehicular Technology, IEEEJournal on Selected Areas in Communications, IEEEINFOCOM, etc.5/21Vol.4, No.5, October 2013

IEEE COMSOC MMTC R-LetterAnalysis of Mobility Impact on the Throughput Performance and MACDesign of Enhancement Schemes in Drive-Thru InternetA short review for “ MAC in Motion: Impact of Mobility on the MAC of Drive-Thru Internet”Edited by Bin LinH. Luan, X. Ling, and X. (Sherman) Shen,, “ in Motion: Impact of Mobility on the MACof Drive-Thru Internet ”, IEEE Transactions on mobile computing, vol. 11, no. 2,pp.305319, Feb. 2012.The demand of high-rate Internet access fromvehicles is ever-increasing due to the fact thatpeople are spending more and more time in theirvehicles currently. However, traditional cellular orsatellite wireless communications can only providelimited available data rate which is far from enoughto deliver the media-rich Internet contents. Drivenby this, the plethora IEEE 802.11b access pointsAPs have been deployed in cities so as to addressthe issue of high-rate cheap Internet access fromvehicles [1][2].802.11 DCF in the newly emerged vehicularenvironment, instead of proposing new MACschemes [4] [5]. The consideration is more practicalfor the plethora IEEE 802.11b APs deploymentsituation.In this paper, the authors contributed at bothperformance evaluation and protocol enhancement.For performance evaluation, the authors provide atheoretical treatment based on a Markov chainmodel which incorporates the mobility of vehiclesin the analysis of DCF. The developed model isaccurate and scalable. It can investigate thethroughput performance under different velocitiesand network scales. The results have shown thatdue to the mobility, the network size of the drivethru Internet is solely dependent on the nodevelocity, which can be applied to optimallyconfigure the DCF by knowing the node velocityonly.Although related performance studies in IEEE802.11 has been well investigated previously, theperformance of IEEE 802.11 in the high-speedlarge-scale drive-thru Internet scenario is still notclear and very challenging. The reasons are twofold. One is that the drive-thru Internet is typicallya much larger network composed of tens orhundreds of users, compared with the small-scaleindoor scenarios. The other is that the IEEE 802.11was originally designed for low-mobility scenarios[3]. Thus, the IEEE 802.11 adopts the contentionbased distributed coordination function (DCF) as itsMAC. In the case of drive-thru Internet, as vehicleshave volatile connectivity due to the fast mobility,whether DCF can fully utilize the cherished accesstime of users and provide them the guaranteedthroughput is questionable.For the protocol enhancement, which is based onthe developed model, the authors propose to furtherenhance the MAC throughput by adaptivelyadjusting the MAC in tune with the node mobility.In particular, the authors propose three assertions asthe guidelines of the DCF design in the highlymobile vehicular environment, and describe theoptimal schemes to determine the channel accessopportunity to fully utilize the transientconnectivity of vehicles. Simulations results haveshown the effectiveness of the scheme to mitigatethe impacts of mobility.To address these issues, the key questions facingthe authors are: how the performance of DCF is inthe high-speed large-scale drive-thru Internet; inwhat fashion the mobility affects the MACthroughput and, more importantly, how to remedythat?In summary, the paper focuses on the DCFperformance by considering high-node mobilities.A systematic and theoretical model is firstdeveloped based on the Markov model to analysisthe quantified impact of mobility (characterized bythe node velocity and moving direction) on theresultant system throughput. Then, protocolenhancement scheme is discussed as the guidelineof the selection of CWmin in different zones to boostthe performance of DCF by accommodating theCompared with reported research, the novelty andchallenge of this paper are that the authors providean elaborate MAC analysis of DCF performance bytaking high-node mobility into account, and modelthe specific DCF in detail, especially, the quantifiedimpacts of mobility on the MAC throughput areshown. In addition, the authors target to proposeenhancement schemes based on the legacy IEEEhttp://committees.comsoc.org/mmc6/21Vol.4, No.5, October 2013

IEEE COMSOC MMTC R-Letterhigh mobility of nodes. Following this researchdirection, future work includes the extension ofthis framework to evaluate the QoS performancefor multimedia applications and the QoS provisionschemes in the high-speed drive-thru Internetscenario.Acknowledgement:This paper is nominated by the MMTC WirelessTechnology for Multimedia Communications(WTMC) Interest Group.References:[5] J. Ott and D. Kutscher, “Drive-ThruInternet: IEEE 802.11b for ‘Automobile’Users,” Proc. IEEE INFOCOM, 2004.[6] V. Bychkovsky, B. Hull, A. Miu, H.Balakrishnan,andS.Madden,“AMeasurement Study of Vehicular InternetAccess Using in SituWi-Fi Networks,”Proc. ACM MobiCom, 2006.[7] L.X. Cai, X. Shen, J.W. Mark, L. Cai, andY. Xiao, “Voice Capacity Analysis ofWLAN with Unbalanced Traffic,” IEEETrans. Vehicular Technology, vol. 55, no. 3,pp. 752-761, May 2006.[8] J. Zhang, Q. Zhang, and W. Jia, “VC-MAC:A Cooperative MAC Protocol in , vol. 58, no. 3, pp. 1561-1571,Mar. 2009.[9] B. Sikdar, “Characterization and Abatementof the Reassociation Overhead in Vehicle toRoadside Networks,” IEEE Trans. Comm.,vol. 58, no. 11, pp. 3296-3304, Nov. 2010.of Electrical and Computer Engineering,University of Waterloo. She has participated andhas been the principle investigator of severalnational and ministry level grants, notably theNational Natural Science Foundation of China(NFSC), the Scientific Research Foundation forthe Returned Overseas Chinese Scholars fromState Education Ministry and Ministry of HumanResources and Social Security, ChinaPostdoctoral Science Foundation, Start-upResearch Fund of Liaoning Province, andFundamental Research Funds for the CentralUniversities in China. Her research tical network integration, networkoptimization and dimensioning, locationproblems, resource allocation, algorithm design,and artificial intelligence.Bin Lin is an AssociateProfessorwiththeDepartmentofInformation Science andTechnology,DalianMaritimeUniversity,Dalian, Liaoning, China.She had received the B.Sc.degreeincomputercommunications and the M.Sc. degree incomputer science from Dalian MaritimeUniversity in 1999 and 2003, respectively, andthe Ph.D. degree in electrical and computerengineering from the University of Waterloo,Waterloo, ON, Canada, in 2009. In 2009, sheheld a postdoctoral position with the 4, No.5, October 2013

IEEE COMSOC MMTC R-LetterImproving the Video Conferencing QoE using the CloudA short review for “Airlift: Video Conferencing as a Cloud Service using Inter-Datacenter Networks”Edited by Jiang ZhuY. Feng, B. Li, and B. Li, “Airlift: Video Conferencing as a Cloud Service using InterDatacenter Networks”, in Proceedings of the 2012 IEEE International Conference onNetwork Protocols (ICNP12), pp. 1-11, 2012.Our way of communication has been fundamentallychanged with the introduction of the Internet, byusing emails, tweets, VoIP, and now videoconferencing. Existing video conferencing serviceslike Skype [1], Google Hangout [2], and iChat [3]have attracted millions of users. Video conferencing,specifically multi-party video conferencing, requireshigh bitrates and low-delay voice/video transmission.With todays' network infrastructure, it is still ing services to consumers through the besteffort Internet.Existing solutions in the literature have traditionallyfocused on the use of peer-to-peer (P2P) [4,5,7] orsimple client-server architectures (e.g., MicrosoftLync). In order to maximize the video bitrate from asingle source to the remaining participants in a videoconferencing application, Steiner tree packing can beemployed for video multicast. However, the problemis NP-complete and, thus, only depth-1 and depth-2trees are generally considered to reduce complexity[4,5]. Another approach is to use intra-sessionnetwork coding as advocated in this paper. With thehelp of network coding, it is possible to formulate theproblem of maximizing the total throughput across allsessions as a linear program, which is easily solvableusing a standard LP solver.A new trend for multi-party video conferencing isoffering it as a cloud-based service. The more popularones, for example, Google and Skype, rely on proxyservers residing inside datacenters that will relay thevideo flows to all other users in the conference [6].The proposed Airlift also takes advantage of thecomputational power of the datacenters as well as thehigh-capacity inter-datacenternetwork.Eachdatacenter is responsible for aggregating incomingvideo flows from users attaching to it, and forwardingthem to the other datacenter, eventually reachingother participants. With such aggregation, the numberof video flows traversing the network is naturallyminimized, without the complexity of tree packingand flow scheduling.http://committees.comsoc.org/mmcThere are a number of important objectives whendesigning a new protocol for multi-party videoconferencing. The first one is performance. It is in theusers' best interest, if the system can provide highervideo bitrates while still maintaining an acceptableend-to-end delay. From the service point of view, thesystem should be simple, and work as a full-servicebroker inside the datacenter. Last but not the least, thesystem should be scalable. In other words, it shouldbe able to handle thousands of flows on one interdatacenter link.With these considerations in mind, the authorsdesigned the Airlift protocol as follows. First, a LPproblem is solved to obtain conceptual flow rates oneach inter-datacenter link with delay bound. Thesolution provides the complete plan to start actualpacket transmission. It is non-trivial to realize thecomplete plan, simply due to the bandwidth overheadand additional decoding delay. The authors proposedto decouple generation (n packets unit) from slidingwindow, so that coded packets are generated within asmall generation (low decoding delay), while asliding window represents all the packets that havebeen sent but not yet acknowledged by the destination.By doing this, the link utilization can be maintainedas high as possible. The final step will be realizingconceptual flows by replicating, splitting, andmerging depending on cases. It is also worth to noticethat each session can be weighted by some factor toachieve basic fairness across participants when theirvideo sources share the bottleneck link.The implementation of Airlift is evaluated withPlanetLab and Amazon EC2 datacenters, where abroker is running as a VM executing network coding,packet processing, and forwarding. A centraloptimizer periodically collects network statistics fromall brokers and refactors the system operation.Experimental results show substantial performanceadvantage over the Celerity [4] system with up to 24xthroughput gain with similar end-to-end delay.8/21Vol.4, No.5, October 2013

IEEE COMSOC MMTC R-LetterThis work is uniquely positioned for high data rateand delay stringent applications with practicalnetwork coding implementation. It does not onlyprovide a promising application layer protocol forhigh quality multi-party video conferencing, but canalso inspire others in the research areas of networkcoding, mobile cloud services, and multimediaadaptations.[16] M. Ponec, S. Sengupta, M. Chen, J. Li and P.Chou,“Multi-ratePeer-to-PeerVideoConferencing: A Distributed Approach usingScalable Coding”, in Proc. IEEE ICME, 2009Jiang Zhu is a seniortechnical leader in EnterpriseNetworking Labs at CiscoSystems, Inc. He has over 15years of industrial experiencebuilding networked systemsranging from embeddeddevices to large-scale mediasystems. His research focuseson adaptive content networking, large-scale datasystems, software defined networking (SDN), cloudservice orchestrations and applications of data miningand machine learning in these fields. He got hisdoctoral degree from Carnegie Mellon Universitywith thesis topic on machine learning and data miningof mobile user behavioral analysis. He also did hisdoctoral study on focusing on SDN and OpenFlow inHigh Performance Networking Group at StanfordUniversity. He also received his M.S. degrees inElectrical Engineering and in Management Science &Engineering from Stanford University. BeforeStanford, he obtained his B.Eng in Automation fromTsinghua University.References:[10] Skype, http://www.skype.com/.[11] Google , https://plus.google.com/.[12] hat.[13] X. Chen, M. Chen, B. Li, Y. Zhao, Y. Wu and J.Li, “Celerity: A Low-Delay Multi-PartyConferencing Solution,” in Proc. ACMMultimedia, pp. 493-502, 2011.[14] C. Liang, M. Zhao and Y. Liu, “OptimalBandwidth Sharing in Multiswarm MultipartyP2P Video Conferencing system,” IEEE/ACMtrans. Networking, vol. 19, no. 6, pp. 1704–1716,Aug. 1996.[15] Y. Xu, C. Yu, J. Li and Y. Liu, “VideoTelephony for End-consumers: Measurementstudy of Google , iChat and Skype,” in Proc.ACM Internet Measurement Conference, pp.371–384, 2012.http://committees.comsoc.org/mmc9/21Vol.4, No.5, October 2013

IEEE COMSOC MMTC R-LetterExploiting Cloud Images for CompressionA short review for “Cloud-Based Image Coding for Mobile Devices – Toward Thousands to One Compression”Edited by Carl James DebonoH. Yue, X. Sun, J. Yang, and F. Wu, “Cloud-Based Image Coding for Mobile Devices –Toward Thousands to One Compression,” IEEE Transactions on Multimedia, vol. 15, no.4, pp. 845-857, June 2013.It is becoming customary to place images on thecloud which has become a huge repository for suchdata. The emergence of cloud images has poseddemands for the development and implementation ofcontent-based image retrieval. Furthermore, this isfacilitating advances, amongst others, on imagecompletion, object recognition, content annotation,and scene modeling. An open question tackled bythis paper is how to exploit cloud images for imagecompression. High compression of images isdesirable for both storage and transmission of thisdata, especially in memory-limited devices andbandwidth-limited networks.The biggest challenge when dealing with cloud-basedimage compression comes from the indefiniteness ofcloud images. Existing compression techniques, suchas those based on prediction and transforms, requirefixed correlated images to compress an image. Thesecorrelated images have to remain unaltered andcannot be deleted once they have been used asreference in compression. However, this requirementis very difficult to guarantee in clouds as the imagescan be changing frequently. New images arebecoming available while others in the cloud can bemodified or deleted in very short periods of time.The authors of this paper propose a method of cloudbased image coding that is different from the methodsthat are currently in use for image coding. Rather thancompressing images in a pixel-by-pixel fashion, thisalgorithm tries to provide a description of the imagesand reconstructs them from available cloud images byexploiting these descriptors. They first utilize theScale-Invariant Feature Transform (SIFT) descriptors[1] to characterize the local features available in animage. Previous work, such as [2] and [3], has

This paper is nominated by the MMTC Wireless Technology for Multimedia Communications (WTMC) Interest Group. References: [1] S. Haykin, “Cooperative radio: Brain-empowered wireless communications,” IEEE Journal on Selected Areas on Communications, vol. 23, no. 2, pp. 201-220, 2005.