Simulating Autonomic SLA Enactment in Clouds Using Case Based Reasoning

With the emergence of Cloud Computing resources of physical machines have to be allocated to virtual machines (VMs) in an ondemand way. However, the efficient allocation of resources like memory, storage or bandwidth to a VM is not a trivial task. On the one hand, the Service Level Agreement (SLA) that defines QoS goals for arbitrary parameters between the Cloud provider and the customer should not be violated. On the other hand, the Cloud providers aim to maximize their profit, where optimizing resource usage is an important part. In this paper we develop a simulation engine that mimics the control cycle of an autonomic manager to evaluate different knowledge management techniques (KM) feasible for efficient resource management and SLA attainment. We especially focus on the use of Case Based Reasoning (CBR) for KM and decision-making. We discuss its suitability for efficiently governing on-demand resource allocation in Cloud infrastructures by evaluating it with the simulation engine.