------------------------------------------------------------------------------

                                       
		     B O S T O N   U N I V E R S I T Y
                                       
			Computer Science Department
                                       
			    C O L L O Q U I U M
                                       
			   Friday, July 11, 1997
                                       
				  3:00 pm
                                       
		 (Coffee served at 2:45 pm, Room MCS 137)

			       Room MCS 135

------------------------------------------------------------------------------


      A SCALABLE CONGESTION CONTROL ALGORITHM FOR INTERNET MULTICAST
                                    
			      Katia Obraczka
		    USC Information Sciences Institute
			       katia@ISI.EDU

     (Joint work with Dante DeLucia, USC Computer Science Department)


  Congestion control is essential for applications to coexist in an
  environment like the Internet.  The vast majority of traffic on the
  Internet today uses TCP's congestion control mechanisms to detect and
  respond to congestion.  However, TCP's point-to-point model does not make
  efficient use of the network when delivering data for group communication
  applications such as information dissemination services and multi-party
  teleconferencing tools: it treats multipoint data delivery as a collection
  of point-to-point flows, thus sending duplicate data repeatedly over the
  same network links.  By using a many-to-many delivery model, multicast
  protocols can greatly improve the efficiency of multipoint data
  distribution.  On the other hand, they can cause considerably more damage
  to the network than point-to-point protocols.  To allow multicast protocols
  to be deployed on the Internet, it is imperative that they incorporate
  mechanisms for handling network congestion.
  
  In this talk, I will present a congestion control mechanism for reliable,
  bulk-data multicast applications.  Our algorithm uses a small set of group
  members, or representatives, to provide timely and accurate feedback on
  behalf of congested subtrees of a multicast distribution tree.  Our
  algorithm does not need to compute round-trip time (RTT) from all receivers
  to the source, nor does it require knowledge of group membership or network
  topology.  Through simulations, we evaluate our algorithm with and without
  TCP cross traffic.  I will present the results of this initial evaluation
  study showing that our algorithm takes advantage of network bandwidth when
  available, yet does not starve competing flows. I will also describe the
  current status of our work as well as talk about directions for future
  research.

Host: Prof. Azer Bestavros (best@cs.bu.edu)

------------------------------------------------------------------------------
For colloquium info, including directions, see http://www.cs.bu.edu/colloquium
     For more information contact Prof. David Yates <djy@cs.bu.edu>
------------------------------------------------------------------------------