On the Complexity of RSSA of Anycast Demands in Spectrally-Spatially Flexible Optical Networks

Abstract

Spectrally-spatially flexible optical networks (SS-FONs) are proposed as a solution to overcome the expected capacity crunch caused by the rapidly growing overall Internet traffic. SS-FONs combine two network technologies, namely, flex-grid optical networks and spatial division multiplexing yielding a significant capacity increase. Moreover, network services applying anycast transmission are gaining popularity. In anycasting, the same content is provided in several geographically spread data centers (DCs), and the requested content is delivered to the network client from themost convenient DC, e.g., minimizing the network traffic and delay. The main optimization challenge in SS-FONs is routing, spectrum and space allocation (RSSA) problem, which can be solved using integer linear programming (ILP). The main goal of this paper is to compare the complexity of various ILP models for routing static anycast traffic in SS-FONs over single-mode fiber bundles (SMFBs). The proposed ILP models apply different modeling techniques, i.e., slice-based and lightpath-based. Moreover, proposed models differ with the core switching (lane changes) capability and consideration of DCs location problem. In order to test the complexity and scalability of models, we run simulations assuming a different number of demands, fibers in SMFBs, candidate paths and DCs.