Priority based multi-stage laxity-aware workload distribution for collaborative vehicular edge computing / Zeseya Sharmin

• 第一著者: Zeseya , Sharmin
• フォーマット: 学位論文
• 出版事項: 2021
• 主題: QA75 Electronic computers. Computer science; T Technology (General)

Technological developments have made it possible for smart machines to access groundbreaking applications. The backend cloud servers are unreliable because of the resultant overhead network to cope with increasing processing requirements. This is usually minimized by the use of edge positions to satisfy the growing demands for computation. A significant function is to maximize productivity and reduce the volume of data transferred into the cloud for fog collection, analysis and storage. Edge locations have commonly been aimed at promoting latency related solutions for end-users. However, resource-restricted regions are frequently flooded by many ongoing demands, whereas the output of delay sensing systems is difficult to sustain at the lowest end-to-end delay. To provide the quality of services (QoS) to resource-limited end-user using computing resources within the data transmission range and to handle the imbalanced workloads because of the traffic density the micro-level fog unit has formed a fog federation in a network that uses underutilized resources to provide service efficiency and achieved energy reduction through this technique by comparing with the traditional non-federated model where multi-access edge computing is used to process data between fog node and the end-users. Most modern innovative vehicular services are delay-sensitive and computationally complex. They pose challenging obstacles for vehicular networks since vehicles are resource stricken with limited computing and storage capacity. Earlier attempts using edge servers get choked up with increasing vehicular traffic. Moreover, workload balancing at available resources, especially there is limited support for priority tasks. In this paper, we propose a collaborative fog computing system to improve balanced fog resource utilization by offloading tasks across the fog federation. The participating fog nodes implement a workload-based offloading decision model, enabling collaboration and suitable fog node selection within the collaborative environment. Furthermore, the system implements a priority-aware multi-queue task scheduling to provide high service throughput. The simulation results demonstrate improved performance for the proposed collaborative fog computing system in terms of queuing delay, delay rate, number of task offloading, and pending tasks.