Robust QoT Assured Resource Allocation in Shared Backup Path Protection Based EONs

TL;DR

This paper presents a robust optimization framework for resource allocation in elastic optical networks that ensures quality-of-transmission under physical impairments and failures, improving survivability and QoT guarantees.

Contribution

It introduces a MILP-based robust optimization model and a novel impairment-aware algorithm with a new sorting technique for enhanced QoT assurance in EONs.

Findings

Achieved around 40% more QoT guaranteed requests in simulations.

First use of bitloading technique for spectrum allocation in survivable EONs.

Proposed SBPP-IA algorithm outperforms existing algorithms in QoT parameters.

Abstract

Survivability is mission-critical for elastic optical networks (EONs) as they are expected to carry an enormous amount of data. In this paper, we consider the problem of designing shared backup path protection (SBPP) based EON that facilitates the minimum quality-of-transmission (QoT) assured allocation against physical layer impairments (PLIs) under any single link/shared risk link group (SRLG) failure for static and dynamic traffic scenarios. In general, the effect of PLIs on lightpath varies based on the location of failure of a link as it introduces different active working and backup paths. To address these issues in the design of SBPP EON, we formulate a mixed integer linear programming (MILP) based robust optimization framework for static traffic with the objective of minimizing overall fragmentation. In this process, we use the efficient bitloading technique for spectrum allocation for the first time in survivable EONs. In addition, we propose a novel SBPP-impairment aware (SBPP-IA) algorithm considering the limitations of MILP for larger networks. For this purpose, we introduce a novel sorting technique named most congested working-least congested backup first (MCW-LCBF) to sort the given set of static requests. Next, we employ our SBPP-IA algorithm for dynamic traffic scenario and compare it with existing algorithms in terms of different QoT parameters. We demonstrated through simulations that our study provides around 40% more QoT guaranteed requests compared to existing ones.