Predicting Resource Allocation and Costs for Business Processes in the Cloud

By moving business processes into the cloud,<br> business partners can benefit from lower costs, more flexibility<br> and greater scalability in terms of resources offered by the<br> cloud providers. In order to execute a process or a part of it, a<br> business process owner selects and leases feasible resources<br> while considering different constraints; e.g., optimizing resource<br> requirements and minimizing their costs. In this context,<br> utilizing information about the process models or the dependencies<br> between tasks can help the owner to better manage<br> leased resources. In this paper, we propose a novel resource<br> allocation technique based on the execution path of the process,<br> used to assist the business process owner in efficiently leasing<br> computing resources. The technique comprises three phases,<br> namely process execution prediction, resource allocation and<br> cost estimation. The first exploits the business process model<br> metrics and attributes in order to predict the process execution<br> and the required resources, while the second utilizes this<br> prediction for efficient allocation of the cloud resources. The<br> final phase estimates and optimizes costs of leased resources by<br> combining different pricing models offered by the provider.