Multicasting in WDM networks with heterogeneous group weights

We address multicasting in a wavelength-routed WDM network in which each destination node has different, as opposed to the traditional uniform, weights for different multicast groups. The weight, in practice, may reflect the popularity of that multicast group at the node. The objective is therefore to serve the multicast groups so as to maximize the total weights of the multicast trees (or equivalently to minimize the weighted overall blocking rate). We propose for this purpose a routing and wavelength assignment (RWA) heuristic that serves the group of the largest weight first (LWF). Since LWF penalizes groups of small weights, we also propose a fairness improvement (FI) heuristic that runs on top of LWF to achieve better fairness among the multicast groups in terms of their respective blocking rates. We show that LWF significantly reduces the weighted overall blocking rate as compared with traditional schemes that do not take group weights into consideration, and FI is effective in improving the blocking fairness. Moreover, we show that FI does not trade off the overall blocking rate for fairness when the group weight heterogeneity is small, the group sizes are small, or the network is densely connected, because of more efficient use of the links. © 2004 Optical Society of America