A New Protection-Based Approach for Link Failure Management of Software-Defined Networks

Abstract

Restoration-based failure management techniques select new paths for software-defined networks (SDN) after diagnosing broken links, which are time-consuming and require steady network inspection. In this paper, a new protection-based link failure management approach is presented for SDN, which attempts to identify links with high probability of failure before they fail definitively. The proposed approach calculates failure index of links by examining three parameters, latency, throughput, and packet loss ratio. In addition, probability density functions (PDFs) of these univariate parameters are used to distinguish between failure-free and broken statuses of a link. Moreover, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">m</i> -sample delay timer technique is applied to adjust the proper trip point and improve detection accuracy of the proposed approach. The results of simulations conducted in Mininet tool have validated effectiveness of the proposed approach in terms of success ratio in finding backup paths, average allocated backup rules, recovery latency, and end-to-end delay. Likewise, influence of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${\boldsymbol{m}}$</tex-math></inline-formula> and Mean Time between Failures (MTBF) on performance of the proposed approach is investigated and the optimal <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${\boldsymbol{m}}$</tex-math></inline-formula> is determined for various circumstances. Eventually, time and message complexity analysis indicates that overhead of the proposed approach is negligible.