The Announcement of

An Invited

Workshop on Middleware

August 28 & 29, 1995

Cambridge, Massachusetts

With the growth of the Internet and the corresponding agreement on a common network and transport infrastructure comes a new opportunity. The consensus on the use of the TCP/IP protocol suite, including protocols for routing, addressing, and forwarding of packets, end-to-end transport mechanisms, and standard protocols for file transfer, remote login, electronic mail and network management has resulted in a wide spread interoperable networking infrastructure. This in turn is creating tremendous opportunities for new and innovative services to be provided over the network. An example of this is the rapid evolution of the World-Wide Web in the last several years.

At the same time, the explosive growth in availability and use of the Internet is creating new challenges. The existing infrastructure must be extended if it is to continue to scale in numbers of users, offer bandwidth guarantees, and support new classes of network applications. These trends demand that we raise the level of common services and introduce new types of higher-level services. Interoperable information management, search, and retrieval mechanisms would create opportunities for new third-party information services. Interoperable payment mechanisms would spur the evolution of electronic commerce. Interoperable multimedia representation formats and exchange protocols would spur a new generation of group support tools. There are opportunities to create many such middleware components, including caching and replication services, indexing services, remote scripting environments, data typing and structuring primitives, and higher level communication abstractions such as multicast and causal broadcast.

The purpose of this workshop is to explore this area of middleware infrastructure. A specific objective is to identify and discuss those areas of services that would sit "above" the traditional network protocols and provide means for extending the the commonly available services on the network to enclose higher layers of abstraction. A critical question to be addressed is the balance between standardization and the exploration of new concepts for middleware services. Thus, identifying the appropriate level for standardization and interoperability is expected to be a major topic of discussion at the workshop. We plan to discuss the vision for where middleware is heading. For example, what kind of middleware support is needed for the integration of databases, information spaces (such as WWW), and personal communication tools. Finally, engineering issues in middleware implementation will be of interest. Questions will be addressed such as how to engineer the interface between middleware and TCP and how to take advantage of underneath IP multicast support in caching and replication services.

The workshop will be organized based on invited participants. A workshop of approximately 75 people is anticipated. Brief (no more than three pages) position papers are welcome to aid in the selection of workshop invitees. These position papers are due no later than 30 April 1995. The workshop itself will last 1.5 days immediately preceding the main SIGCOMM'95 conference.

Position papers should be sent by electronic mail in ascii to

Questions should be addressed to the Workshop Program Chair, Dr. Barry M. Leiner,

Workshop Program Committee: