earticle

영어

Providing an efficient and reliable multicast for data dissemination applications on a large scale is a challenge, especially when the applications require a very short delivery delay and high throughput. The combination of a local recovery approach based on active services with those using FEC/ARQ gives rise to a new class of reliable multicast protocols called APES "Active Parity Encoding Services". This paper carries out a comparative study between reliable multicast protocols belonging to this class in terms of loss recovery latency. We use the simulation to study the impact of coding/decoding time and loss rate on the performance of this class of protocols. Our numerical results show that the approach Get-Repairs Store-Repairs, besides the reduction in storage space at the active router, provides a substantial gain in terms of loss recovery latency and thus contributes to improve the real time reliable multicast.