Abstract

Overlay multicast has been recently proposed as alternative scheme to traditional IP multicast. However, since data packets are transmitted by relaying between group members in a form of unicast packet, overlay multicast has larger link stress that indicates the number of times that the same data traverses a particular physical link than IP multicast. Larger link stress may cause packet loss by network and node congestion. More serious problem is that packet loss on one group member can affect all descendents' packet reception. Thus, it is very important to construct another type of overlay multicast tree to minimize packet loss. In this paper, we propose to build a new overlay multicast tree, BASE (Bandwidth Aware overlay multicaSt architecturE). Especially, BASE uses available bandwidth metric instead of hop count or delay in order to construct overlay multicast tree. The first feature of BASE is to reduce packet loss probability by locating group member with more available bandwidth at upper level so that packet loss seldom happens at upper position. The second feature of BASE is to immediately resolve both network and node congestion according to varied traffic load by designating proper number of children dynamically. Finally, with simulation results, we demonstrate the advantage of BASE protocol in heavy traffic load or other simulation environments. Even though heavy traffic load is employed, only BASE shows the most acceptable achievements.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call