Implementation of an Energy-Efficient VM Management Strategy in the Openstack Cloud Platform

The objective of this thesis is to implement a set of mechanisms within the Openstack platform to assign virtual machines (VMs) to physical machines (PMs) while minimizing energy costs and the cost of licenses assigned to PMs. The project must adapt to Aruba’s infrastructure, so an analysis of its characteristics has been conducted, and two implementation proposals have been evaluated: an ad hoc solution and a solution that integrates Openstack’s Watcher component. Watcher is an optimization component capable of executing various algorithms in parallel, called strategies. After choosing the solution with Watcher, the algorithms that can schedule VMs in PMs have been analyzed, specifically comparing an optimal mathematical model (developed for another thesis) and a heuristic algorithm. Both algorithms have been implemented within the Watcher component. All implementation challenges have been studied and resolved by developing some independent strategies. To facilitate the development and testing phase, a component that can simulate Openstack in a completely transparent manner for Watcher has been implemented.