Smart Edge Service Update Scheduler : An Industrial Use Case

Abstract

Software systems need to be maintained and frequently updated to provide the best possible service to the end-users. However, updates sometimes, cause the system or part of it to restart and disconnect, causing downtime and potentially reducing the quality of service. In this work we studied and analyzed the case of a large Nordic company running a service-oriented system running on edge nodes, and providing services to 270K IoT devices. To update the system while minimizing downtime, we develop a smart edge service update scheduler for a service-oriented architecture, which suggests the best possible update schedule that minimizes the loss of connections for IoT devices. Our approach was validated by applying the scheduling algorithm to the whole system counting 270k edge nodes distributed among 800 locations. By taking into account the topology of the software system and its real-time utilization, it is possible to optimize the updates in a way that substantially minimizes downtime.