Heurística online para aquisição de instâncias em cloud pública: um olhar sobre a demanda de e-commerce.

Abstract

Large companies are spending a lot on Infrastructure as a Service (IaaS) and it is growing. Public cloud providers offer several ways to purchase their services. These options are usually divided into paying later for the time you use the service or paying in advance, with an associated discount, for a reservation of the service to be used in the future. There are many studies that seek to minimize the cost of allocating resources in IaaS providers, based on the estimate of future demand and the purchase options made available by the providers. One of the possible solutions is the use of online algorithms that are able to decide when to reserve during the arrival of demand. This study seeks to evaluate the performance of an online heuristic taking into account the demand of a large e-commerce company.