Self-configuration model based neural predition and agent technology for cloud infrastructure

Cloud computing including hardware, software and networks are growing towards an ever increasing scale and heterogeneity, becoming overly complex. In order to manage such growing complexity and information overload, nowadays Cloud researchers propose to enrich the Cloud with Autonomic Computing properties. In fact the autonomic computing focuses on self-adaptable computing systems to the maximum extent possible without human intervention or guidance. The Cloud Infrastructure is partially automated. This limitation results from the absence of an architectural model for describing distributed resources and their interdependencies in terms of configuration. This article proposes a model of distributed resources in the form of a set of resources interconnected and automated in order to predict and configure cloud resources. We propose to use the agent technology in order to provide a distributed solution for the decentralized resources self-configuring and the adaptation of the Cloud infrastructure. We enhanced the proposed model with neural prediction method and broker agent in order to ameliorate the performance of the Cloud Infrastructure. Moreover, the proposed model has been validated using a formal policy.