An Architecture-Based Autonomous Engine for Services Configuration and Deployment in Hybrid Clouds

摘要

As cloud computing offers scalability, extensibility, elasticity, flexibility and cost savings to the customers of cloud service providers, there is a growing trend towards migrating services to the cloud. Hybrid clouds, which comprise nodes both in the private cloud and in the public cloud, have emerged as a new model for service providers to deploy their services. However, to deploy services in hybrid clouds is a complex task as services are in essence distributed applications. What is more, there is heterogeneity among hybrid clouds. This article proposes an autonomous engine for services configuration and deployment in hybrid clouds. The automation is enabled by the definition of generic information model, which describes all the information relevant to the deployment and configuration of services with the same abstractions, including the required resources, service dependencies and business objectives. In addition, to shield the heterogeneity of hybrid clouds, we define mapping rules for model transformation. We also deploy a three-layer architecture application on Openstack and CloudStack to validate the correctness of our approach.