Temporal Computing Resource Allocation Scheme With End Device Assistance

Abstract

With the development of the Industrial Internet of Things (IIoT) and edge computing technology, edge servers have been used to allocate computing resources to IIoT devices for alleviating the huge computing needs of IIoT applications. Aiming at the problems of unbalanced distribution and inefficient use of computing resources during computing resource allocation, we propose a temporal computing resource allocation scheme with end device assistance. First, by establishing the temporal interest degree which uses the long-term interest accumulation and short-term interest change, IIoT devices are divided into different interest queues, the end devices with similarly temporal interest to the resource request device are sensed. Second, a social RippleNet based on a temporal knowledge graph is constructed. By analyzing the temporal characteristics and propagation process of social attributes in RippleNet, the temporally social similarity between the end devices and the resource request device is extracted, then the final computing resource allocation decision is generated by using the computing resource selection algorithm we proposed based on temporally social similarity. The simulation results show that the scheme we proposed can effectively improve the efficiency of computing resource allocation and balance the distribution of computing resources in the IIoT environment.