Middleware support for scalable services

Abstract

Services targeted to the home environment must deal with software and hardware heterogeneity of client and server machines. Distributed object based middleware is attractive for building distributed services in such heterogeneous environments for a number of reasons. In particular, distributed objects can hide the complexities that arise from distribution and heterogeneity in the underlying system. To deal with the scale of home based applications and to reduce the effects of high latencies in such environments, we have been exploring efficient implementations of distributed object systems that exploit caching and replication of service state. In this paper, we explore issues related to object state caching and communication for distributed object systems.