Dealing with reliable event-based communications by means of layered multicast

Abstract

Nowadays, it is demanding to have multicast services able to jointly provide a high degree of reliability and good performances despite of failure occurrences. However, the current practice to tolerate failures in such services mainly consists in the adoption of retransmission-based methods, which obtain reliable multicast delivery with severe performance fluctuations and instability. An alternative solution is to apply coding to proactively recover the lost packets by reconstructing them from what has been received. Coding-based approaches achieve more predictable and stable performances, since the delivery time does not depend on the loss patterns imposed by network dynamics. However, they have been scarcely adopted within the context of large-scale multicast services due to important issues related to their tuning. In this paper we have gleaned from the experience of Layered Multicast, commonly applied to multimedia content delivery, and extended it to address the tunability problem of coding-based recovery approaches in large-scale multicast services.