Route Partitioning Scheme for Elastic Optical Networks With Hitless Defragmentation

Abstract

Hitless defragmentation has been introduced as an approach to limit the spectrum fragmentation in elastic optical networks without traffic disruption. It facilitates the accommodation of new requests by creating large spectrum blocks as it moves active lightpaths (retuning) to fill in gaps left in the spectrum by expired ones. Nevertheless, hitless defragmentation witnesses limitations for gradual retuning with the conventionally used first-fit allocation. The first-fit allocation stacks all lightpaths to the lower end of the spectrum. This leads to a large number of lightpaths that need to be retuned and are subject to interfering with each other's retuning. This paper proposes a route partitioning scheme for hitless defragmentation in order to increase the admissible traffic in elastic optical networks. The proposed scheme uses route partitioning with the first-last fit allocation to increase the possibilities of lightpath retuning by avoiding the retuning interference among lightpaths. The first-last fit allocation is used to set a bipartition with one partition allocated with the first fit and the second with the last fit. Lightpaths that are allocated on different partitions cannot interfere with each other. Thus, the route partitioning avoids interference among lightpaths when retuning. We define the route partitioning problem as an optimization problem to minimize the total interference. We then introduce an integer linear programming problem (ILP) that yields an optimal routing and partitioning. We prove that the route partitioning problem is non-deterministic polynomial-time hard (NP-hard). We present a heuristic algorithm for large networks, where the ILP used to represent the route partitioning is not tractable. The simulation results show that the proposed route partitioning scheme using the first-last fit outperforms the conventional first-fit allocation for hitless defragmentation in term of allowable traffic.