Towards a Fuzzy Bigraphical Multi Agent System for Cloud of Clouds Elasticity Management

Abstract

A major ingredient that encourages the wide adoption of Cloud Computing solution is their ability to allocate and des-allocate resources in automatic and optimal manner according to customers' needs or what is referred to Cloud Elasticity. Indeed, Cloud Elasticity is influenced by different dependable factors, principally: unpredictable variable workload, Cloud resource limitation, possible system failures, etc. Thus, ensuring Cloud elastic behavior of systems is a very challenging task. Our first contribution is to propose a Multi Agent System for Cloud of Clouds Elasticity Management (MAS-C2EM). The interacting agents cooperate to ensure Cloud Elasticity through a set of strategic solutions at both micro (Intra Cloud Infrastructure/Intra Cloud Software) and macro (Inter Cloud) levels. We define the Inter Cloud Elasticity strategy using a novel Top Clouds Cooperation solution. The given strategy is based on the fuzzification of Pareto dominance mechanism. In the second contribution, we adopt Bigraphical Reactive System (BRS) as an effective formalism to specify the MAS-C2EM at both structural and behavioral aspects, resulting in the definition of Bigraphical Multi Agent System for Cloud of Clouds Elasticity Management (BigMAS-C2EM) semantical model. Besides, we formally analyze the model via some inherent properties using the model checker BigMC.