Cooperative offloading of tasks with deadlines in mobile cloud computing

Mobile cloud computing (MCC) offers an agile solution that facilitates mobile applications to exploit cloud computing technology. Compared to offloading tasks to a remote cloud through wide area network (WAN), offloading to nearby cloudlets via WLAN links offers much smaller network delay. Resource capacity at cloudlets may become a bottleneck upon peak demands. A remote cloud then constitutes an appropriate back-up despite longer network delay. In this work, we design a cooperative task offloading scheme for tasks with hard deadlines, utilizing both local cloudlets and a remote cloud. By leveraging the compact exponential optimization technique, we design an online algorithm that simultaneously guarantees system efficiency, truthfulness, and social welfare maximization. A moderate parameterized competitive ratio is proven, and simulation studies reveal the algorithm's close-to-offline-optimum performance under practical settings.