A Scalable Middleware Solution for Fast Data Transport

Multipoint distribution of data is a requirement for various types of applications in such areas as distance learning, multimedia conferencing and group collaboration. The economy of scale of multicast transport for multipoint distribution is not in doubt. The challenge is to develop technologies that will satisfy the varied requirements of these applications. However, IP Multicast has proven difficult to deploy on a large scale, and it can not satisfy the requirements of the plurality of multicast applications. Many multicast protocols have been proposed as alternatives to the open service model of IP Multicast. But no one set of protocols is capable of satisfying the various, often at odds, service requirements of multicast applications with heterogeneous multicast receivers. This paper proposes a middleware approach to provide core protocols, functions and interfaces, the combination of which can be selected and customized to meet the specific demands of various multicast-based applications. At its core, the proposed middleware solution will allow the routing of both datagrams and streams, using multiple channels and variants of the standard transport protocols, over clustered trees that scale well for large number of hosts in a multicast group. Key-Words: Data transportation, multimedia streaming, high performance data transfer, middleware, routing Proceedings of the 5th WSEAS Int. Conf. on DISTANCE LEARNING AND WEB ENGINEERING, Corfu, Greece, August 23-25, 2005 (pp128-133)