A distributed self-balancing policy for virtual machine management in cloud datacenters

Cloud Computing is a crucial computational paradigm for modern companies because it can discharge them from managing their ever growing IT infrastructure. Dynamically offering a plenty of computational resources, the cloud can also simplify the execution of CPU-intensive applications. Modern data centers for cloud computing are facing the challenge of a growing complexity due to the increasing number of users and their augmenting resource requests. A lot of efforts are now concentrated on providing the cloud infrastructure with autonomic behavior, so that it can take decisions about virtual machine (VM) management across the datacenter's nodes without human intervention. While the major part of these solutions is intrinsically centralized and suffers of scalability and reliability problems, we investigate the possibility to provide the cloud with a decentralized self-organizing behavior. To this purpose we present a novel VM migration policy suitable for a distributed environment, where hosts can exchange status information with each other according to a predefined protocol. The main goal of the policy is to balance the computational load on datacenter's physical hosts by conveniently moving virtual machines (VMs). We tested the policy performance by means of an ad hoc built simulator.