Gateway Nodes Selection Strategy for Hierarchical Edge-Cloud

Edge-cloud has been viewed as a promising approach to effectively reduce communication latency by pushing computing servers to the edge of network. A distributed online task scheduling approach for hierarchical edge-cloud has been proposed in prior work. However, this approach hasn't fully addressed load balancing issue of the system. In this paper, I model the hierarchical edge-cloud as a multi-queue system, and propose a gateway nodes selection strategy based on greedy algorithm to dynamically balance loads among those queues. Simulation result demonstrates the proposed strategy outperforms baseline algorithms in term of failed requests rate and percent idle time.