Multi-granularity and robust grooming in power- and port-cost-efficient IP over WDM networks

Previous studies on the power efficiency or port savings in IP over WDM networks have required explicit knowledge of the traffic between each network node pair and the resource assignment of each fiber link. However, it is difficult to estimate this information accurately in live networks, due to the inherent features of IP traffic, e.g., bursts, unpredictability, and variability. Furthermore, the granularity of the demands on an IP-level connection tends to be diverse, and the number of ports consumed in Optical Cross-Connects (OXCs) tends to grow due to the use of hybrid grooming (i.e., traffic grooming with an optical bypass) for power savings. Therefore, it is critical to achieve both power efficiency and port savings in realistic IP over WDM networks. In this paper, we investigate the problem of multi-granularity and robust grooming for power- and port-cost-efficient IP over WDM networks. First, a key parameter, the Multi-Granularity Power Ratio (MGPR), which is the sum of the different single-granularity power ratios, is proposed. Second, the MMPR (Minimizing Multi-granularity Power Ratio) method is used to compute the Traffic Distribution Vectors. Finally, we present the multi-granularity and robust grooming approach, called the Maximizing Hop First (MXHF) approach, where hybrid grooming is adopted to improve power efficiency and waveband merging is utilized to reduce the port cost. Simulation results demonstrate that the MXHF approach can achieve power efficiency and port savings comparing with state-of-the-art robust grooming methods.