Online Partitioning for Scalable and Survivable Optical Networks

Given the trend of rapid growth, the manageability and survivability of future optical networks, even for a single administrative domain, are imminent challenges. For a large-scale mesh network with dynamic traffic, maintaining the global state information in a centralized fashion is impractical. Hence, distributed schemes are needed to organize nodes and to manage state information in a more localized manner. One such scheme for organizing nodes is to logically partition or cluster the nodes into a hierarchical structure. However, in optical networks, physically disjoint routes are sought for failure-protection purposes, where the working path and the backup path of a request traverse diverse shared risk link groups (SRLGs). Hence, the clustering of nodes should also be effective in yielding SRLG diversity. In this paper, we address the problem of determining the appropriate clustering of nodes for a large wavelength division multiplexed (WDM) network. We suggest an easy-to-implement, dynamic, and distributed technique that forms clusters adaptively in response to the current network conditions, e.g., node connectivity, bandwidth availability (or traffic load), and SRLGs. We find that considering risk group sharing in network partitioning best handles both scalability and survivability. Simulation results verify the viability of the proposed approach.