Predição de cargas para provisionamento de recursos em computação em nuvem

Abstract

Since the outbreak of the Internet in the 90s, the area of computing has experimented great changes, accompanied with all the needs for a better experience of its users. In this context, humanity has became aware of new possibilities to expand their business. Thus, Cloud Computing came with the purpose of providing computing resources as services, cheapening the cost of purchase of products related to Information Technology. Cloud computing is an attractive computing model that allows resources to be provided according to the demand, i.e., users can rent the resources from the cloud as necessary. Though cloud computing brings many benefits to its users, the construction and maintenance of large data centers show up expensive. The high consumption of electricity is among the main concerns about the costs, that besides spending also damages the environment. To alleviate these problems, came up a model called Green Computing, which has as one of its objectives the proposed deployment of data centers that consume less power and generate less pollution. Based on this context, this paper proposes an algorithm that will facilitate the management systems decisions in data centers, concerning to the shutdown of underutilized servers and other decisions that aim to meet the Green Computing.