随机可靠组播协议的改进  

Improved Reliable Multicast Transport Protocol

作　　者：姜腊林[1] 吴敏[1] 

机构地区：[1]长沙理工大学计算机与通信工程学院,湖南长沙410114

出　　处：《湖南大学学报（自然科学版）》2011年第5期79-82,共4页Journal of Hunan University:Natural Sciences

基　　金：湖南省科技厅资助科研项目(2010FJ3115)

摘　　要：差错控制是保证组播可靠性的最基本技术,合理的缓存管理可以提高差错控制的效率.在随机可靠组播协议(RRMP)的基础上提出了改进的可靠组播协议(IRMP).IRMP使用短期缓存满足本地恢复需要,使用长期缓存满足远程恢复需要.在本地恢复时,利用组播发送重传请求,利用随机后退算法减少重复报文.在远程恢复时,使用指定修复节点(DRN)恢复丢失的报文.为了避免因DRN退出或故障导致长期缓存报文的丢失,提出了DRN转移的方法.在NS2平台上对IRMP和RRMP进行仿真的结果表明,IRMP有效地降低了缓存需求,减少了恢复时延,具有较高可扩展性.Error control is a basic technology to ensure multicast reliability.Reasonable buffer management can improve error control efficiency.This paper has put forward an improved reliable multicast protocol（IRMP） on the basis of randomized reliable multicast protocol（RRMP）.IRMP used a short buffer to meet local recovery requirement,and used a long buffer to meet remote recovery requirement.In local recovery,the retransmission request was sent by multicast,and the randomized back-off algorithm was used to reduce repeat message.In remote recovery,a designated repair node（DRN） was responsible for the repair.In order to avoid the quit or default of DRN,this paper has put forward a method for DRN transfer.Simulation results for IMRP and RRMP in NS2 have shown that IMRP has good ability to reduce the buffer requirements and recovery latency,and also improves its scalability.

关 键 词：组播 可靠组播 缓存管理 差错恢复 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]