Analyzing Two Different Objectives of the WDM Lightpath Reconfiguration Problem

Lightpath Reconfiguration is a common task that is often performed in order to improve resources utilization. The WDM reconfiguration problem becomes non-trivial when the new set of lightpaths requires the release of resources previously seized by the old ones, but, in order to assure continuous traffic flow, the old ones cannot be torn down before the new ones are set up. Under these conditions the reconfiguration is in a deadlock state. This can only be solved by temporary disruption of some connections. Depending on the network operator’s policy, it could be desirable to either: a) minimize the total number of disrupted connections or, b) minimize the number of concurrent disrupted connections (at any time during the reconfiguration process). Considering either of the objectives in the lightpath reconfiguration problem bring NP-hard problems on. In this paper, we, firstly, show that both objectives are conflictive. We propose two algorithms for solving the problem considering the second objectives in two steps. We perform simulations that shows the trade-off between these two objectives.