Abstract

Middleware for supporting the programming of distributed systems has been developed since almost the beginnings of the existence of the Internet. Different approaches have been implemented in order to support different ways of communication, architecture, and data sharing. In this paper we present a schema called MatchMaker developed for developing applications requiring a replicated architecture, dynamic and partial synchronization. Two implementations of MatchMaker are presented: one based on Java's RMI distributed objects for synchronizing applications running inside a local area network or networks without proxies, fire-walls or NAT addresses, and the other for applications in a more restrictive environment. MatchMaker has been successfully used for developing many applications supporting collaborative distributed learning. ©2006 IEEE.