A scalable reliable multicast transport protocol with advanced buffer management

Many reliable multicast transport protocols have been proposed to achieve efficient scalabilities and reliabilities in the field of multicast transmission. Tree-based reliable multicast schemes are one of those protocols, which divide the multicast tree into sub-trees and allocate a single node ( known as repair node ) in each sub-tree to do the task of loss packet recovery. The repair node, in each sub-tree is used to buffer and retransmit the loss packets. There is a great deal of problems in buffering packets waiting for a long time until they get positive acknowledgments from all children receivers of the repair node’s sub-tree. The problem gets even worse when the number of children receivers under the repair node increase over a certain limit or during heavy transmission. In that case, a buffer overflow will certainly occur which creates network congestion, also the throughput, scalability and in general, the performance of the system will be greatly decreased. This work introduces a new strategy based on distributing the burden of packets buffering and retransmission on a number of selected receivers (SRs) in each local group instead of entrusting this mission to only one node, the repair node. The proposed protocol, which is called, ‘The Selected Receiver Reliable Multicast Transport Protocol (or abbreviated to SRRMTP),is capable of solving the problems of buffering these packets waiting for positive acknowledgements. This distribution of the packets, can help in solving the congestion problem, increasing the network throughput, and decreasing the number of packets retransmitted from the source. Moreover, it can decrease the stability and latency time served, as well as distributing the burden of the resending process thereby increasing the protocol scalability. The proposed protocol is implemented over two environments; the first is the fixed network, and the second is the combined(fixed and mobile)network environments. Simulation using (C++) was performed to study and compare the performance of the proposed protocol with the previous protocols that have been based on the repair node strategy. The results showed that the performances of the proposed protocol are better than the repair node protocols in terms of through put ,packet stability time and latency time, and packets retransmitted from the original sender. The proposed protocol was also more scalable than the repair node protocols in the case of the increment of the number of nodes in each local region or The results showed that the performances of the proposed protocol are better than the repair node protocols in terms of through put,packet stability time and latency time,and packets retransmitted from the original sender. The proposed protocol was also more scalable than the repair node protocols in the case of the increment of the number of nodes in each local region or with an increment of the number of local regions in the multicast session.

Preview Text FK 2016 49 - IR.pdf Download (2MB) | Preview

Actions (login required)

View Item