JOB SERVICE DEMAND BASED RESOURCE ALLOCATION POLICY FOR CLOUD-FOG COMPUTING ENVIRONMENTS

Abstract

In the development of 5G mobile network, Fog computing becomes an emerging component of the Cloud computing paradigm as the 5G core to assure the diverse computational demands of IoT applications can be satisfied. A cloud-based application requires a combined use of cloud and local resources for its processing. Resource allocation for cloud-based jobs plays an important role to achieve both QoS and efficient resource utilization of a Cloud-based computing environment. This study considers a hierarchical computing architecture in 5G network made up of various computational machines from user device, edge server, fog to cloud center. The system is to handle diverse IoT applications in heterogeneous computational resources. This study proposes a resource allocation policy that takes account of job characteristics in terms of job service demands with the aim at improving job service quality. We develop simulation software to investigate the proposed policy. Numerical results point out that the proposal reduces the average response time by 5% to 9% and yields better resource utilizations in comparison to a best-effort Round Robin policy.