Revenue-driven service provisioning for resource sharing in mobile cloud computing

Abstract

A new mobile cloud architecture has recently been proposed, where neighbouring mobile devices are brought together as a cohort for resource sharing. Thus, a mobile application can be partitioned into multiple tasks and be performed on different mobile devices, without offloading some tasks to remote cloud services for their processing. In this work, we consider a broker-based architecture, where a proper reward mechanism is used to incentivize users to share their available resources as a service with other mobile devices. Within such a system, each mobile device is rational and justified in maximizing its revenue by using its relatively limited resources to complete the requests allocated by the broker. How to select the appropriate service requests from all incoming requests and complete the selected requests on time so that the revenue can be maximized has become a critical issue for resource sharing. To address this issue, we propose a joint resource sharing and request scheduling approach called RESP (REvenue-driven Service Provision for mobile devices) in a move towards a lightweight one-phase approach for handling request selection, request scheduling and resource allocation for mobile devices. We evaluate the performance of our proposed algorithm through a number of experiments and the experimental results validate the efficacy of our approach.