Thomas Rausch, Philipp Raith, Padmanabhan Pillai, Scharam Dustar
This work was presented at the 4th ACM/IEEE Symposium on Edge Computing and was awarded with best demo award. Congrats to the team led by Thomas Rausch.
We present an end-to-end system for operating energy-aware cloudlets with a low-footprint cluster manager and an adaptive client-side load balancing approach. Our system is designed for small-scale high-density compute clusters that host stateless services and have stringent energy resource constraints. It features cluster and service management, runtime monitoring, adaptive load balancing and cluster reconfiguration policies. Furthermore, we present an experimentation and analytics system that allows coordinated execution of complex workload experiments to evaluate different operational strategies.