From: Susumu Harada (harada@cs.washington.edu)
Date: Wed Nov 10 2004 - 02:23:16 PST
"The PIM Architecture for Wide-Area Multicast Routing"
S. Deering et al.
This paper proposes a new method for implementing multicast routing
tailored for sparse distribution of clients over a large network such as
the Internet, called the Protocol Independent Multicast (PIM)
architecture. The paper first identifies the shortcomings of the
traditional IP multicast model and the various protocols for managing
group membership and distribution routes. It then proposes their
solution, PIM, which combines some of the features of these other
protocols and offers the ability to choose between different types of
multicast trees.
One major shortcoming of the paper was their lack of evaluation of their
proposed architecture. They seem to have empirical evidence for the CBT
protocol and why it does not scale, but they do not provide any results
for their own method. The list of outstanding issues at the end of the
paper also seem to suggest that the idea is still in its incubation phase
and has not been put through any rigorous experimentation. The only
indication of any experimentation is in the conclusion, in which they
vaguely state that "simulation and implementation efforts conducted
characterize configuration criteria and deployment issues". Therefore, it
is suspect whether their proposed architecture really offers "robustness,
flexibility, and scaling properties" as stated in the abstract..
Given that this paper was written in 1995, I would like to know how much
of it is actually in place in today's multicasting protocols. The
contribution of the paper would have been greater had it included
empirical data on how well it actually works in realistic conditions. As
multimedia content on the Internet continue to increase and the bandwidth
of end hosts also grows, the ability to multicast efficiently and robustly
should become an increasingly important issue.
This archive was generated by hypermail 2.1.6 : Wed Nov 10 2004 - 02:23:16 PST