Self-optimising Decentralised Service Placement in Heterogeneous Cloud Federation

Clouds have been relevant business drivers and computationalbackends for a wide range of applications, includingIoT, e-health, data analytics. To match the complex needs ofsuch comprehensive set of different kinds of applications, inrecent times there is an emerging need for new paradigmsand forms of Clouds, organised according to a federated, heterogeneousand distributed structure. To exploit heterogeneityand localisation, in order to enhance the overall performances,ensure energy efficiency, reduce costs for resource providers andin the meantime enhance the user experience, proper serviceplacement solutions are required. However, conducting efficientdeployments in such a scenario is complex due to the dynamicnature of applications, resources, users. As a consequence, therethe a need for scalable, distributed, adaptive, context-awaresolutions characterised by high-efficiency and reduced overhead.We propose a highly distributed, self-adaptive solution aimed atoptimising the overall deployment of cloud services by means ofpoint-to-point interactions occurring among clouds and cloudletsbelonging to the same federation. The contribution of this paperis the definition of a service exchange mechanism, its Markovchainbased modelling and thorough experimental evaluation.