Intelligent computational offloading for mobile-edge server computing and hybrid optimal resource allocation
K. Muralidhar,S. Siva Shankar,Bhuvan Unhelkar,Tulika Chakrabarti,Prasun Chakrabarti
DOI: https://doi.org/10.1007/s11042-024-18117-y
IF: 2.577
2024-02-05
Multimedia Tools and Applications
Abstract:To get beyond the limitations of mobile devices, the mobile cloud is an emerging technology. Offloading resource-intensive applications to distant data centres enabled by the cloud is helping us achieve that. Real-time mobile user applications suffer while using a remote computing solution since MDs encounter increased network response times and delays. In this paper, we suggest an intelligent computational offloading model for MEC. The suggested model aims to apply the DL technique, which automatically chooses the computing source depending on performance, energy consumption, and workload. These factors are used to select the best edge server. For this, a Modified LSTM is proposed in this work. Additionally, TS on the edge cloud infrastructure are directly impacted by VM availability; as a consequence, VM availability is calculated while managing TS. When allocating resources, VM calculations like Make span, task completion times, resource consumption, and migration costs are considered. The capacity to deliver functioning services in the required amount of time after the task is offloaded to the VM is considered while allocating VM resources. Given that this is an optimization problem, the NBUJ hybrid optimization algorithm is developed to solve this issue. At last, the performance of the developed model is validated with the existing models in terms of fitness, migration cost, makespan, resource utilization and so on. And, it is noted that the developed model attains 97.4% of accuracy and 0.0274% of FNR which validates the performance of the proposed model.
computer science, information systems, theory & methods,engineering, electrical & electronic, software engineering