The CloudMIG Approach: Model-Based Migration of Software Systems to Cloud-Optimized Applications

Cloud computing provides means for reducing overand under-provisioning through enabling a highly flexible resource allocation. Running an existing software system on a cloud computing basis can involve extensive reengineering activities during the migration. To reduce the correspondent effort, it is often possible to deploy an existing system widely unmodified in IaaS VM instances. However, this simplistic migration approach does not solve the challenge of overand under-provisioning or scalability issues per se, as our experiments using Eucalyptus and the popular open source system Apache OFBiz show. Moreover, current migration approaches suffer from several further shortcomings. For example, they are often limited to specific cloud environments or do not provide automated support for the alignment with a cloud environment. We present our model-based approach CloudMIG which addresses these shortcomings. It aims at supporting SaaS providers to semi-automatically migrate existing enterprise software systems to scalable and resource-efficient PaaS and IaaS-based applications. To facilitate reasoning about the suitability of certain cloud environments for a given system and the degree of alignment during the reengineering process, we introduce the Cloud Suitability and Alignment (CSA) hierarchy. For example, Apache OFBiz used in our experiments is initially categorized “cloud compatible” but not “cloud optimized” as it does not exploit the cloud’s advantages. Keywords-Approach CloudMIG, Cloud Computing, Modelbased software migration to cloud-based applications, Resourceefficient cloud-based applications, Eucalyptus, CSA hierarchy.