Experimental evaluation of the source specific multicast model in mobile environments

Although the single source multicast model is quite mature, the implementation of its protocols became available only very recently and experimental results to evaluate this model are still missing, specially in mobile environments. Hence, our goal is to experimentally evaluate the single source multicast model against the two multicast mobility proposals of the IETF: remote subscription and bi-directional tunneling techniques. We experimentally confirm that the group management query, in remote subscription, and the router advertisement interval, in bi-directional tunneling are unacceptably high and not adjusted for mobile environments. The handover latency of both techniques is closely related with those intervals. Other results show that both techniques have similar throughput and packet loss in home network, having the remote subscription higher throughput in remote access networks. This throughput decrease in remote access networks reveals one of the drawbacks of the bidirectional tunneling technique: the non-optimal routing. Its was also concluded that RFC 3810 is not prepared to deal with mobile environments