Understanding the Design Tradeoffs for Cooperative Streaming Multicast
Understanding the design tradeoffs for cooperative streaming multicast Animesh Nandi‡⋄, Bobby Bhattacharjee†, Peter Druschel⋄ ⋄MPI-SWS ‡Rice University †University of Maryland Technical Report MPI-SWS-2009-002 April 2009 ABSTRACT fers from previous works that have compared CEM design choices Video streaming over the Internet is rapidly increasing in popular- qualitatively [1, 21] or analytically [42, 44, 7, 22], and with those ity, but the availability and quality of the content is limited by the that have compared specific CEM protocols empirically [26, 45]. high bandwidth cost for server-based solutions. Cooperative end- It is not our intent to recommend any single approach or pro- system multicast (CEM) has emerged as a promising paradigm for tocol. Instead, we explore the CEM design space, cleanly identify content distribution in the Internet, because the bandwidth over- the tradeoffs that apply to these systems, tease out different compo- head of disseminating content is shared among the participants of nents that are responsible for different aspects of observed behav- the CEM overlay network. Several CEM systems have been pro- ior, and partition deployment scenarios into regions where different posed and deployed, but the tradeoffs inherent in the different de- systems excel. signs are not well understood. A systematic comparison of CEM systems is non-trivial. These In this work, we provide a common framework in which different systems deliver data over a diversity of data topologies (tree, multi- CEM design choices can be empirically and systematically evalu- tree, mesh, and hybrids) which are constructed and maintained us- ated. Our results show that all CEM protocols are inherently lim- ing different control and transport protocols.
[Show full text]