Architectures for Hardware and Software Scalable Multiwavelength Networks

Abstract

The multiwavelength multihop lightwave network has recently emerged as a viable candidate for a new wide-area telecommunications infrastructure capable of bringing the benefits of high speed multimedia services to large and small users alike. In this paper, we consider the performance of one such network, known as the recursive grid. A unique aspect of this network is its scalability, meaning that as the network grows larger, both the per-user access hardware, and the per-user computational complexity needed to set up new user-to-user end connections, remain constant. Compromised is the overall network utilization efficiency, since sub-optimum routing is always involved. Thus, this network invokes the complexity management principle: when capacity is over-abundant, compromise utilization efficiency for software simplicity. Extensive simulation studies of this network were undertaken for a wide variety of network parameters and a non-uniform population-driven traffic model, and an enormous volume of results have been generated. Based on these results, our overall conclusion is that the recursive grid approach does indeed yield both scalable traffic management algorithms and good utilization efficiency of the network optical resources. Moreover, it appears that the approach effectively exploits the traffic nonuniformities such that the overall efficiency diminishes slowly, if at all, as the number of access stations grows large, re-confirming that the reconfigurable WDM approach based on a scalable architecture is indeed a strong candidate for the future telecom infrastructure.