A residue-based approach for resource provisioning by horizontal scaling across heterogeneous clouds

A cloud's elasticity which makes it attractive is limited to its physical resources. As a measure to prevent this limitation from being a hindrance to the attraction, the cloud inter-operates with other clouds. This improves its resource strength by borrowing necessary resources from the clouds with which it inter-operates. Normally, this borrowing happens from one cloud to satisfy one request arriving to it, thereby enhancing its elasticity to a certain extent. Further enhancement is achieved by horizontal scaling across clouds to satisfy a single request. In this work, the enhancement is taken one step further to propose a novel residue-based resource provisioning technique for a cloud to perform horizontal scaling by splitting a request dynamically based on the resource availability in itself and in the clouds with which it inter-operates. Also, a greedy technique is presented to rank clouds and hence, additionally contribute to the enhancement of elasticity and, in turn, to the effectiveness of the resource provisioning technique used in inter-operating cloud environments. Results obtained from real-time heterogeneous cloud environments prove the effectiveness of the proposed algorithms.