routing for finding the shortest path between two nodes, such as Intermediate System to
Intermediate System (IS-IS) protocol [63] and Open Shortest Path First (OSPF) protocol
[49]. The tree information such as forming BNRs, down-BNRs, up-BNR, nodes cost are kept
at each node dynamically. At each BNR and LMR, the possible processing includes:
i) Perform a route table look up.
ii) Partition the set of destinations based on their next hops.
iii) Replicate the packets for each of next hops.
iv) Modify the list of destinations in each of copies.
For Xcast and Xcast+, the processing will be made for multicast packets. For SEM and
SReM, the processing will be made for multicast packets and control messages1. The
performance of these protocols will highly depend on their respective processing.
To evaluate and compare the four algorithms considered in this thesis, three metrics are
considered, which are:
-I- The first one is processing cost (at nodes), including costs due to header encoding /
decoding, exchanging of control messages. The processing cost is occurred at nodes
when processing being made for multicast packets / control messages.
-I- The second one is delivering cost (over links), including costs in delivery of multicast
packets and control messages. The delivering cost is the one occurred at delivering
multicast packets / exchanging control messages.
-I- The third metrics is join/leave cost. The join/leave cost is an alternative measure of
latency to reflect the real delay in join/ leave processing.
These metrics are given in more details as follows.
The following assumptions and definitions are used to calculate the total processing /
delivering cost in a multicast tree.
1 In THIS simulation, HBH and REUNITE have not been chosen to compare to our proposed protocol SReM.
This is because in both HBH and REUNITE the join/leave messages are sent periodically , which results in
some unfairness and inconvenience to compare with the Xcast, Xcast+, SEM and our proposed SReM.
83