Comparative Analysis on Cluster-based Algorithms for Failure Management in Optical Networks

The occurrence of failures in transparent optical networks (TONs) potentially compromises their reliability, leading to loss of information or even link disruption. In that regard, as the conventional methods for failure management become difficult with the increasing complexity of TONs, the use of machine learning algorithms has emerged as an alternative strategy given their sole capabilities concerning network scalability and self-awareness. Therefore, in this paper, the failure management performance of three cluster-based algorithms (k-means, fuzzy c-means, and Gaussian mixture models) are compared on a real-world testbed in terms of failure classification errors.