9508274 Towsley Wide-area networks pose new communication-related challenges for multimedia applications as packet loss, longer delays, and larger user populations become important factors. While emerging ATM networks provide communication services for supporting such multimedia applications today's internetworked environment and legacy make end-to-end ATM connectivity between two end hosts unlikely in the foreseeable future. Research in this project represents a timely and important step in developing the network infrastructure needed to support real-time multimedia applications. This proposal includes a mix of theoretical and experimental research aimed at the design and evaluation of scalable services for supporting real-time multimedia applications in a wide-area network or internetwork. Specifically, the researchers will: 1) develop and evaluate receiver-oriented approaches for providing scalable reliable multicast services. The fundamental concerns are to understand how to allocate protocol functionality among the sender and receivers, how to use multicast to provide reliable multicast, and how to use application-specific semantic information in error recovery protocols; and 2) develop and evaluate end host mechanisms, such as playout adjustment algorithms and special-purpose error-recovery protocols, for enhancing the quality of service attained by several different real-time applications. The researchers will exploit application-specific information about the component applications - in this case their different timing constraints. *** ******* Darleen Fisher Program Manager Networking and Communications Research Program 4201 Wilson Blvd. Room 1175 Arlington Va 22230 (703) 306-1949 (703) 306-0621 (FAX)