Exploração do Espaço de Projetos de Sistemas Heterogêneos Aplicada ao Problema de Alocação de Recursos em Nuvem

Abstract

Cloud computing services offer on demand computing resources to applications. Finding the best cloud resource allocation that fits application budget, meets performance, and follows the applications constraints is still a research challenge. Our studies have shown that such challenge is quite akin to the Design Space Exploration (DSE) problem once that they both have to find suitable hardware configurations in a large design space, having incompatible objectives subject to several constraints. In that scenario, the goal of this work is to solve the cloud resource allocation problem by adopting techniques that originally are applied to solve the design space exploration problem. This work designed and develop a software extension named MultiExplorer-VM from a DSE tool called MultiExplorer. This new extension has a workflow to work with the cloud resource allocation problem by applying techniques originally targeted to solve the DSE problem. A comprehensive set of experiments has been performed to statistical validation and to compare solutions from our proposal to other existing research work, focused on the cloud resource allocation problem, based on the Paramount Interation (PI) technique. The results show that the proposed extension achieves significant (better) results compared to the PI technique. According to the model objectives and constraints, for some applications, solutions from MultiExplorer-VM are similar to the optimal cloud configuration.