Resource Allocation Method for Fairness and Efficient Utilization of Network and Computational Resources in Edge Networks

Abstract

AbstractCross-domain data fusion is becoming a key driver of the growth of numerous and diverse applications in the IoT era. We have previously proposed a new cross-domain data fusion platform, the Geo-Centric Information Platform (GCIP), which enables IoT data fusion in a geolocation-based edge network. GCIP dynamically produces Spatio-Temporal Content (STC) by combining cross-domain data in each geographic area and then delivers them to users. However, when a large amount of IoT data is required for STC creation, there is a heavy load on the GCIP network and computational resources. This paper introduces a network-wide pre-processing method. When multiple flows with different loads on network and computational resources arrive at an edge network, (1) throughput degradation (efficiency issues) and (2) inequity in resource allocation (fairness issues) may occur. In this paper, we propose a comprehensive resource allocation method for efficient and fair utilization of network and computational resources. Through the numerical verification, we have demonstrated that the proposed method successfully improves efficiency and fairness by 26% and 38%, respectively.