Abstract: In this paper we are concerned with disseminating high-volume data streams to many simultaneous applications over a low-bandwidth wireless mesh network. For bandwidth efficiency, we propose a \textitgroup-aware stream filtering approach, used in conjunction with multicasting, that exploits two overlooked, yet important, properties of these applications: 1) many applications can tolerate some degree of ``slack'' in their data quality requirements, and 2) there may exist multiple subsets of the source data satisfying the quality needs of an application. We can thus choose the ``best alternative'' subset for each application to maximize the data overlap within the group to best benefit from multicasting. An evaluation of our prototype implementation shows that group-aware data filtering can save bandwidth with low CPU overhead. We also analyze the key factors that affect its performance, based on testing with heterogeneous filtering requirements.
Keywords: distributed computing
PDF (590K)(Dartmouth only)
Copyright © 2008 by Taylor & Francis.The copy made available here is the authors' version; for a definitive copy see the publisher's version described above.