A reliable dissemination protocol for interactive collaborative applications
Proceedings of the third ACM international conference on Multimedia
Log-based receiver-reliable multicast for distributed interactive simulation
SIGCOMM '95 Proceedings of the conference on Applications, technologies, architectures, and protocols for computer communication
A Comparison of Known Classes of Reliable Multicast Protocols
ICNP '96 Proceedings of the 1996 International Conference on Network Protocols (ICNP '96)
RMTP: a reliable multicast transport protocol
INFOCOM'96 Proceedings of the Fifteenth annual joint conference of the IEEE computer and communications societies conference on The conference on computer communications - Volume 3
A comparison of sender-initiated and receiver-initiated reliable multicast protocols
IEEE Journal on Selected Areas in Communications
Design a hierarchical cache system for effective loss recovery in reliable multicast
APPT'07 Proceedings of the 7th international conference on Advanced parallel processing technologies
A reliable overlay video transport protocol for multicast agents in wireless mesh networks
International Journal of Communication Systems
| Hi-index | 0.00 |
Server—based local recovery for reliable multicast will perform efficiently only when sufficient processing and buffering resources are available at the servers. In this paper we examine the buffer requirements of servers for reliable multicast. We show how the server buffer requirements depend upon the packet arrival process at the server and the duration of time during which a packet needs to be held in its buffer. This latter quantity depends upon the required number of retransmissions and the length of time between retransmissions of the packet. We find that buffer requirements are very sensitive to small values of the number of retransmissions of a packet within the packet inter-arrival time but increase very slowly with the number of receivers a server is responsible for supplying repairs. We also determine the effect of the number of servers and buffer size at a server on the end—system throughput and network bandwidth usage.We examine three buffer replacement policies; FIFO, FIFO with minimum buffer holding time or FIFO—MH (that we propose) and LRU. Based on our performance study and noting that it could also be easily implemented, we recommend the use of the simple FIFO buffer replacement policy.