Computation offloading is a key concept in Mobile Cloud Computing: it concerns the capability of moving application components from a mobile device to the cloud. This technique, in general, improves the efficiency of a system, although sometimes it can lead to a performance degradation. To decide when and what to offload, we propose the use of a method for determining an optimal infinite scheduler, which is able to manage the resource assignment of components with the aim of improving the system efficiency in terms of battery consumption and time. In particular, in this paper we define a cost/reward horizon method for Mobile Cloud Computing systems specified in the language MobiCa. By means of the model checker UPPAAL, we synthesize an optimal infinite scheduler for a given system specification. We assess our approach through a case study, which highlights the importance of a scheduler for reducing energy consumption and improving system performance.
A Cost/Reward Method for Optimal Infinite Scheduling in Mobile Cloud Computing
MORICHETTA, ANDREA
;
2016-01-01
Abstract
Computation offloading is a key concept in Mobile Cloud Computing: it concerns the capability of moving application components from a mobile device to the cloud. This technique, in general, improves the efficiency of a system, although sometimes it can lead to a performance degradation. To decide when and what to offload, we propose the use of a method for determining an optimal infinite scheduler, which is able to manage the resource assignment of components with the aim of improving the system efficiency in terms of battery consumption and time. In particular, in this paper we define a cost/reward horizon method for Mobile Cloud Computing systems specified in the language MobiCa. By means of the model checker UPPAAL, we synthesize an optimal infinite scheduler for a given system specification. We assess our approach through a case study, which highlights the importance of a scheduler for reducing energy consumption and improving system performance.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.