Recently, businesses using the Internet have proliferated, and a demand has emerged for technology that offers stable service quality regardless of unexpected load or disturbances, by assigning resources to services on demand has emerged. In a traditional study of resource allocation, improvement of an existing method has been performed (Hashimoto et al., 2006), but sufficient study has not been done to date from the perspective of adding value to the supplied services. In this paper, the objectives and challenges of service operators are investigated and a method is proposed that is the allocation of highly available resources for on demand data center services.