Inteligência computacional como suporte à tomada de decisão no gerenciamento de capacidade de serviços de computação em nuvem

Abstract

Cloud computing services have changed people’s lifestyles and the use of cloud computing has increased due to its convenience, possibility of cost reduction and wide vaiety of integrated applications. On the other hand, companies that provide this type of service need to meet the growing demand with high quality of service and manage capacity in line with the strategy in a faster and more objective way. The alternative to help management is the use of computational intelligence and therefore the present work aims to propose and test a solution through a system that predicts the use of server resources and the method used is the test and evaluation of regression and prediction models applied to a set of artificially generated sequential data based on a literature search that identifies the characteristics of these data. As a result, a neural network without feedback produced a model that has a great fit to the test data of the peak daily use of the processor and it is concluded that managers can use the forecast data of processor use as a tool in making management decisions of capacity.