No abstract

No abstract

Repeated access to WWW pages currently makes inefficient use of available network bandwidth. A Distribution Point Model is proposed where large and relatively static sets of pages (e.g. magazines or other such media) are distributed via bulk multicast to LAN distribution points for local access. Some access control issues are discussed. Hopwise Reliable Multicast (HRM) is proposed to simplify reliable multicast of non real time bulk data between LANs. HRM uses TCP for reliability and flow control on a hop by hop basis throughout a multicast distribution tree created by today's Internet MBone.

No abstract

Background and Use of Distributed Computer Services. .