DEOMAC—Decentralized and Energy-Efficient Framework for Offloading Mobile Applications to Clouds

Abstract

AbstractAlthough smart mobile devices are virtuous and aggrandize, they are still valued as circular computing devices. To augment the capabilities of smart devices, mobile cloud computing (MCC) utilized offloading techniques. Computational offloading overcomes the circulation of SMD by offloading the intensive work to other server systems with enhanced performance and resources. In earlier research, offloading frameworks require modification in binary executable or special compilation during offloading, and not considered dynamic QoS parameter which is a major countenance for authentic-time applications. Some of the frameworks have not fixated on how virtual machines (VMs) are allocated to execute a compute-intensive task. A dynamic and energy-efficient offloading framework is required to amend the battery consumption of mobile contrivances with minimum response time by acclimating fluctuating network conditions efficaciously. A Congruous mechanism is required for the allocation of resources at the cloud in a decentralized manner. The primary goal of this research is to design and implement an energy efficient offloading framework which ameliorates the capabilities of mobile contrivances with minimum battery consumption and response time by considering dynamic QoS parameters for offloading decision. This research presents a Decentralized and energy efficient offloading framework which ameliorates the energy capability of mobile contrivances by offloading computation rich task to other servers. An offloading decision model is proposed utilizing neuro-fuzzy controller (NFC) to prognosticate the execution environment of a task either at mobile or at cloud. It additionally provides decentralized resource allocation of Virtual Machines in Cloud with the utilization of adaptive cuckoo search optimization algorithm (ACSOA). The proposed framework is implemented in authentic-time environment utilizing Android platform and battery performances of mobile contrivances are evaluated utilizing AccuBattery application. The proposed framework was evaluated by varying the number of images in the database to prove that the proposed method is better than subsisting methods. After that, the proposed framework is compared with existing frameworks to prove that response time and energy consumption are better than the traditional frameworks.KeywordsOffloadingEnergy efficientDecentralizedResource allocationDFOMACACSOA