Towards Custom Cloud Services

In today’s highly dynamic economy, businesses have to adapt quickly to market changes, be it customer, competition- or regulation-driven. Cloud computing promises to be a solution to the ever changing computing demand of businesses. Current SaaS, PaaS and IaaS services are often found to be too inflexible to meet the diverse customer requirements regarding service composition and Quality-of-Service. We therefore propose an ontology-based optimization framework allowing Cloud providers to find the best suiting resource composition based on an abstract request for a custom service. Our contribution is three-fold. First, we describe an OWL/SWRL based ontology framework for describing resources (hard- and software) along with their dependencies, interoperability constraints and meta information. Second, we provide an algorithm that makes use of some reasoning queries to derive a graph over all feasible resource compositions based on the abstract request. Third, we show how the graph can be transformed into an integer program, allowing to find the optimal solution from a profit maximizing perspective.