Towards an approach for cloud service composition in Multi-Cloud environment based QoS using deep Q-learning

Cloud Computing is internet-based computing where resources are offered over the internet on-demand and based on Pay-as-you-go. The advantages of Cloud computing have encouraged many enterprises to migrate to it and adapt its services. Sometimes users’ requests are not met from only one cloud and to better satisfy users’ complex requirements in front of the huge number of cloud services, multiple services from the various cloud can be combined by using Cloud Service Composition (SC). SC is an NP-Hard problem divided into services selection and service composition. Few works have combined these two sub-problem for service composition in a Multi-cloud environment. In this paper, we propose an approach to Service Composition in a Multi-Cloud environment based on cooperative Agents to merge these two sub-problems. Such us, our approach combines DQL for obtaining the optimal composition with locale selection of atomic service for each abstract service.