Distributed System Support for Voice in Cedar
Douglas B. Terry
Daniel C. Swinehart
Polle T. Zellweger
Computer Science Laboratory
Xerox Palo Alto Research Center
Abstract: The CSL Voice project at the Xerox Palo Alto Research Center is an effort to define and validate experimentally an architecture that can incorporate live and recorded voice into a networked computing environment. We have developed a prototype voice system consisting of Etherphones, a Voice Control Server, and a Voice File Server, in addition to personal workstations. An Etherphone is a special processor that connects to a telephone and transmits encrypted, digitized voice directly over an ethernet. Etherphones and workstations communicate with the Voice Control Server using a remote procedure call (RPC) protocol to establish voice conversations. To record or playback voice, conversations can be set up directly between an Etherphone and the Voice File Server. Programmer interfaces on the workstations allow voice to be edited and easily integrated into Cedar applications. Work on the Voice Project has allowed a variety of distributed systems issues to be explored such as real-time communication protocols, the suitability of RPC for complex control, services for heterogeneous clients, dynamic binding, security, managing large shared data, and garbage collection.