Publication Phare-Lip6

[ Publication ]

Improving Feedback Merging for Source-Adaptive Layered Multicast Schemes
To appear in Kluwer Cluster Computing Journal
Rezende José F., Duarte Otto Carlos Muniz Ba and Pujolle Guy
March, 2003
Anglais - 24 p - Abstract
Current source-adaptive layered multicast schemes exploit merging capabilities at special nodes in the network to combine feedback from the multicast group. The merging procedures reduce network load and avoid feedback implosion at the source. In this paper, we demonstrate how to increase the efficiency of a merger node. We show how to tune two mechanisms: ``feedback transmissionNULLNULL, on the timer settings used by receivers, and ``temporal mergingNULLNULL, on the timer settings used by intermediate nodes. Our approach is generic, relating purely to timer settings, and does not touch on ``content mergingNULLNULL, or the decision regarding which packets to forward. We investigate the effects of temporal merging on packet suppression. We show that periodic feedback transmission coupled with a suitable timer setting at the merger node increases efficiency.