SIMULADOR PARA ASIGNACIÓN DE RECURSOS EN NUBES BASADAS EN TEORÍA DE JUEGOS Y SISTEMAS DE REPUTACIÓN

Abstract

The present work aims to design a model based on fundamental game theory and reputation systems concepts that help to improve resource allocation mechanisms in Cloud Computing and significantly reduces SLA violations, by the inclusion of Cloud provider reputation as a new parameter in the negotiation process between client and provider; therefore this process was modeled as a non-cooperative game, in which Cloud providers compete for resource allocation of cloud user´s tasks, and bid their offers in order from lowest to highest reputation, these offers will be processed by an intermediary agent, who will use the beta reputation mechanism in order to select the cloud provider that best suits the user's selection criteria. In order to perform the experiments, a cloudsimbased simulator will be developed in a web environment using the JSF framework and PrimeFaces library components to display the results graphically and simplify the user interfaces creation.