Improving Container Deployment in Edge Computing Using the Infrastructure Aware Scheduling Algorithm

The increase of applications with low latency requirements has promoted edge computing as an enabler technology by bringing computational resources closer to end-users. However, this new paradigm presents several challenges, such as the fast and continuous provision of applications on geographically distributed heterogeneous devices at the edge, often with constraint resources. Currently, there are several strategies for scheduling applications in edge environments to decrease application deployment time. However, they do not consider the network bandwidth available or the download queue on each node. In this paper, we present a new scheduling strategy called Infrastructure Aware considering these characteristics. We validated our proposal through simulation. Results show that the Infrastructure Aware scheduling algorithm can, on average, decrease the deployment latency by more than 52% and 40% compared to the Kube-Scheduler and Layer Locality strategies, respectively.