Portal de Periódicos da CAPES

Workload Scheduling in Vehicular Networks With Edge Cloud Capabilities

2019; Institute of Electrical and Electronics Engineers; Volume: 68; Issue: 9 Linguagem: Inglês

10.1109/tvt.2019.2927634

1939-9359

Ibrahim Sorkhoh, Dariush Ebrahimi, Ribal Atallah, Chadi Assi,

Privacy-Preserving Technologies in Data

In order to support the development of 5G technologies, researchers are actively engaged in addressing the challenges accompanying the emerging 5G applications. Unquestionably, an eminent technology gaining significant research attention is edge computing. Vehicular edge computing brings data storage and computing capabilities as well as hosting support applications that comprise emerging vehicular services and applications which demand low-delay processing, to the edge closer to the vehicles, reducing response time and increasing reliability, therefore achieving the holistic vision of the tactile Internet. In this context, this paper considers a vehicular network with edge computing capabilities deployed at road side units, and addresses the problem of workload offloading as well as scheduling of computation tasks on the computing resources available at the edge. The challenge here is the high mobility of the vehicles and hence their short residence time within the coverage range of the road side units hosting the edge computing resources. A joint problem considering the communication and computation resources, as well as the latency requirements of the workload is formulated and the scheduling is shown to be NP-Hard. Subsequently, efficient solutions based on Lagrangian relaxation are derived and presented. We evaluate numerically the proposed methods and show their closeness to the optimal solutions.