Traffic-based reconfiguration for logical topologies in large-scale WDM optical networks

Wavelength-division multiplexing (WDM) technology has emerged as a promising technology for backbone networks. The optical layer based on WDM technology provides optical routing services to the upper layers such as the packet-switching layer and the time-division multiplexing (TDM) layer over the generalized multiprotocol label-switching (GMPLS) paradigm. The set of all-optical communication channels (lightpaths) in the optical layer defines the logical topology for the upper layer applications. Since the traffic demand of upper layer applications fluctuates from time to time, it is required to reconfigure the underlying logical topology in the optical layer accordingly. However, the reconfiguration for the logical topology is reluctantly disruptive to the network since some lightpaths should be torn down and some traffic has to be buffered or rerouted during the reconfiguration process. Therefore, it needs to have an efficient transition method to shift the current logical topology to the new one so as to minimize the effect of the reconfiguration on the upper layer traffic. This paper proposes several heuristic algorithms that move the current logical topology efficiently to the given target logical topology in large-scale wavelength-routed optical networks. In the proposed algorithms, the performance improvement/degradation of data transmission [transmission delay or distance between a source-destination (s-d) pair] caused by a new lightpath is considered as benefit for establishing the new lightpath. The proposed algorithms construct the new logical topology starting from a lightpath with the largest benefit to the user traffic. Simulation experiments have been performed to evaluate the proposed algorithms in comparison with existing algorithms in a National Science Foundation Network (NSFNET)-like network model with 16 nodes and 25 links. The results show that the proposed algorithms yield much better performance (shorter average packet hot distance) than previous algorithms mostly with comparable computation time.