A metaheuristic approach for source traffic grooming in elastic optical networks

Abstract

The increasing demands for higher spectrum efficiency have made elastic optical networks (EON) to be an interesting area for the research community. The possibility that some network parameters are changed according to different types of services allows that optical paths can be established using just enough spectrum as needed. Since individual optical paths have to be separated by guard bands, a large part of the available spectrum is wasted. The efficient utilization of available network capacity is among network operators' major concerns and therefore, traffic engineering techniques that provide additional capacity on the existing infrastructure are of the interest. As such, the traffic grooming technique has been analyzed in this article. By applying the grooming technique, the basic problem in EONs, namely routing and spectrum allocation is extended and made more complex, especially in the case of realistic networks. Compared to electrical traffic grooming applied in traditional wavelength division multiplexing networks, traffic grooming directly at the optical layer is an attractive option for EON because it provides the opportunity to groom several same source connections into one transmitter and in such a way eliminating the guard bands. In this article, we solve the grooming problem in EON with static traffic demands using the bee colony optimization metaheuristic approach. The objective is to serve all traffic demands such that spectrum savings are maximized and total spectrum usage is minimized. The results indicate considerable spectrum savings compared to the scenario without grooming and a significant improvement considering some referenced grooming algorithms.

Preservation of shared backup path ink‐node content based elastic optical datacenter networks

“…The EONs, also called as flexi grid optical networks can offer high scalability needed for the high capacity and high speed data center communication. 10 Here the concept of frequency slot is used, which indicates the spectrum width of a channel. As illustrated in Figure 2, the spectrum is partitioned into frequency chunks, increasing the channel density, enabling higher number of connection requests that can be transferred by a channel.…”