An SRN-based resiliency quantification approach

TitleAn SRN-based resiliency quantification approach
Publication TypeConference Paper
Year of Publication2015
AuthorsD Bruneo, F Longo, M Scarpa, A Puliafito, R Ghosh, and KS Trivedi
Conference NameLecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Date Published01/2015

Resiliency is often considered as a synonym for faulttolerance and reliability/availability. We start from a different definition of resiliency as the ability to deliver services when encountering unexpected changes. Semantics of change is of extreme importance in order to accurately capture the real behavior of a system. We propose a resiliency analysis technique based on stochastic reward nets that allows the modeler: (1) to reuse an already existing dependability or performance model for a specific system with minimal modifications, and (2) to adapt the given model for specific change semantics. To automate the model analysis an algorithm is designed and the modeler is provided with a formalism that corresponds to the semantics. Our algorithm and approach is implemented to demonstrate the proposed resiliency quantification approach. Finally, we discuss the differences between our approach and an alternative technique based on deterministic and stochastic Petri nets and highlight the advantages of the proposed approach in terms of semantics specification.