Coordination Pattern-Based Approach for Auto-Scaling in Multi-clouds

Today, cloud providers cover a huge variety of optimized offers. Many of them are similar, but in detail there exist subtle differences. An example is the automatic scaleout of computing resources if the traffic increases, in order to achieve high availability and fault tolerance. The cloud also supports scaling-in if the load falls again, in order to save costs. The paper aims to abstract such offerings in form of coordination patterns that capture the functional and nonfunctional semantics of the required services in a cloud-agnostic way. Concrete pattern solutions can be deployed on any cloud that provides the respective services, and also be run using a platform-independent middleware.