Modeling, evaluation and provisioning of elastic service-based business processes in the cloud. (Modélisation, évaluation et mise en oeuvre de l'élasticité des applications à base de services dans le cloud)

Cloud computing is being increasingly used for deploying and executing business processes and particularly Service-based Business Processes (SBPs). Among other properties, Cloud environments provide elasticity at different scopes. The principle of elasticity is to ensure the provisioning of necessary and sufficient resources such that a Cloud service continues running smoothly even when the number or quantity of its utilization scales up or down, thereby avoiding under-utilization and over-utilization of resources. It is obvious that provisioning of elastic infrastructures and/or platforms is not sufficient to provide elasticity of deployed business processes. In fact, it is also necessary to consider the elasticity at the application scope. This allows the adaptation of deployed applications during their execution according to demands variation. Therefore, business processes should be provided with elasticity mechanisms allowing their adaptation to the workload changes while ensuring the desired functional and non-functional properties. In our work, we were interested in providing a holistic approach for modeling, evaluating and provisioning of elastic SBPs in the Cloud. We started by proposing a formal model for SBPs elasticity. To do this, we modeled SBPs using Petri nets and defined two elasticity operations (duplication / consolidation). In addition, we proposed to intertwine these elasticity operations with an elasticity controller that monitors SBPs execution, analyzes monitoring information and executes the appropriate elasticity operation (duplication/consolidation) in order to enforce the elasticity of SBPs. After facing the challenge of defining a model and mechanisms for SBPs elasticity, we were interested in the evaluation of elasticity before implementing it in real environments. To this end, we proposed to use our elasticity controller as a framework for the validation and evaluation of elasticity using verification and simulation techniques. Finally, we were interested in the provisioning of elasticity mechanisms for SBPs in real Cloud environments. For this aim, we proposed two approaches. The first approach packages non-elastic SBPs in micro-containers, extended with our elasticity mechanisms, before deploying them in Cloud infrastructures. The second approach integrates our elasticity controller in an autonomic infrastructure to dynamically add elasticity facilities to SBPs deployed on Cloud platforms