Analysis of Offloading Decision Making in Mobile Cloud Computing
Huaming Wu · Universitätsbibliothek der FU Berlin Hochschulschriftenstelle u. Dokumentenserver · 2015
Besides lightweight Internet applications, there is an increasing demand from mobile users for computation-heavy and energy-hungry applications that are being deployed to mobile devices. Running complex applications on such devices is however challenging due to the strict constraints on their resources, e.g. limited computational capacity, battery lifetime and network connectivity. Offloading computation-intensive parts of mobile applications to a capable cloud server is an effective way to alleviate a tussle between resource- constrained mobile devices and resource-hungry mobile applications, and thus boosting the device’s performance. Offloading decisions may involve multiple factors such as resource and component availability, connectivity intermittence and network capacity. Potential benefits obtained from mobile cloud offloading include time and energy saving, which can be achieved by deciding what, where, how and when to offload correctly. Unlike previous works that only focus on some specific issues in the offloading process, our time- and energy-aware offloading decision process is based on multiple perspectives. For instance, what defines the name of the candidate tasks to be offloaded through application partitioning; where describes the type of surrogate and choosing the appropriate offloading target (e.g. local, cloudlet and cloud) in which the application has to be offloaded; how introduces offloading plans that enable the device to schedule offloading operations; when considers the offloading conditions and dynamic changes of context, since sometimes offloading may not be worthwhile at all. This work covers both the theoretical and practical sides of offloading decision making. Mathematical and queueing models are applied and evaluated by numerical simulations and real world experiments. Specifically, the contributions of this thesis can be summarized as follows: • Studying how to effectively and dynamically partition a given application into local and remote parts while keeping the total cost as small as possible. • Proposing static and dynamic methods based on multi- criteria decision making, in order to find out what resource is the most appropriate one for offloading. • Comparing different offloading operations of a mobile terminal equipped with multiple radio access technologies (e.g. 3G, LTE, Bluetooth, WLAN) and determining how to leverage the complementary strength of WLAN and cellular networks by choosing heterogeneous wireless interfaces for offloading. • Exploring the energy-delay tradeoffs for different types of applications (e.g. delay-tolerant and delay-sensitive applications) based on some combined metrics.