Workload-aware Management Targeting Multi-Gateway Internet-of-Things
Ioannis Galanis, Theodoros Marinakis, Iraklis Anagnostopoulos · 2019
Edge Computing has risen as a new promising computing paradigm, that relies on connected devices to communicate and exchange data closer to the end-user. Specifically, gateways are a critical component as they receive requests from multiple edge devices and perform data processing. However, the continuously increased number of edge devices that each gateway serves and the growing network traffic result in (1) unbalanced workload execution of applications offloaded on the gateway; and (2) unequal distribution of workloads in multi-gateway environments. In this paper, we propose (1) an intra-gateway methodology to detect and balance application slowdown in terms of local unevenly distributed progress; and (2) a distributed inter-gateway methodology to balance workload among multiple gateways, in order to achieve a global progress threshold.