Evaluation of MPLS P2MP distribution tree algorithms

Multiprotocol Label Switching (MPLS) Point to Multipoint (P2MP) can be used as an underlying technology for transmitting video data. Reducing the transmission delay on a network is very important to maintain the quality of service for video transmission. Therefore, the hop count of the path between a source and a destination node should be reduced using MPLS P2MP routing technology. We considered the routing algorithm of MPLS P2MP from the aspects of the hop count and the sum of link lengths. We evaluated three algorithms on these aspects with simulations: the conventional shortest path tree (SPT), the minimum cost paths heuristic (MPH), and the constrained MPH (CMPH), which is MPH under hop constraint. We propose CMPH as an example of a dynamic Steiner tree algorithm with hop constraint. The simulations showed that CMPH is a well balanced algorithm satisfying the need for a small average hop count and small sum of link lengths at the same time. We also used the ratio of destination nodes within hop constraint as a scale for evaluation. In this scale, CMPH is the best algorithm among the three.