Authors:
Carolos Livadas,
Idit Keidar, and
Nancy Lynch.
Fast abstract in the International Conference on Dependable Systems and
Networks (DSN'01), pages B44-B45, July 2001.
Abstract:
With the increasing use of the Internet, multi-party communication and
collaboration applications are becoming mainstream. This trend calls
for high-performance multicast services that scale to large groups and
higher bandwidth requirements. Although packet loss characteristics
have a large impact on the performance of multicast services, few
protocols attempt to adapt and actively exploit such
characteristics. In this paper, we describe a reliable multicast
protocol that exploits packet loss locality through caching. Several
studies have observed that packet losses in multicast communication
are bursty, i.e., links drop numerous multicast packets while
temporarily congested. Thus, consecutive losses as witnessed by
individual hosts are likely to occur on the same lossy link. By
caching pertinent information regarding the error recovery of prior
losses and optimistically presuming that future losses occur on the
link responsible for prior losses, our protocol streamlines the
recovery of future losses. This scheme demonstrates how packet loss
locality can be actively used to reduce the recovery latency and the
bandwidth overhead of multicast error control. Moreover, in view of
increasing our confidence in the correctness and performance of our
protocol, we use a rigorous design approach.
Download paper:
ps,
ps.gz,
pdf.
Last modified: Mon Jul 1 14:36:07 EDT 2002