Review: SRM

Design of reliable multicast protocol is driven by varying requirement for
reliability by different application. But it is very hard to come up with
a single multicast model that optimally meets the functionality,
scalability and efficiency requirement of all these applications. The
focus of this paper is to present a framework that meets minimal
definition of reliable multicast and leave the rest of functionality and
flexibility to applications. The framework named ‘Scalable Reliable
Multicast’ (SRM) purposed in this paper relies on the underlying IP
multicast network and requires no explicit support from the network to
support reliable multicast delivery. Author has used a distributed
whiteboard sharing application to illustrate the use of SRM on a global
scale with varying no of users.

In order to provide reliable delivery SRM requires a naming scheme in
which data has unique and persistent names. Whenever an application
generates a new data, it is multicast to the groups and each group member
is individually responsible for detecting loss and requesting
transmissions. In order to improve the efficiency in the presence of data
loss, SRM is based on a loss recovery mechanism that requires a member of
group that detects loss, to send a multicast request for retransmission of
lost data. This prohibits other members who have lost the same data from
sending the same request reducing duplicate retransmission requests for
same data. To evaluate the efficiency of this loss recovery mechanism,
author has presented its working for different network topologies.

This paper discusses different protocols to be used in SRM with the focus
on SRM’s request and repair algorithms. Author has used simulation to
prove the efficiency of SRM’s working.
Received on Tue Oct 31 2006 - 07:59:52 EST