Energy Efficient Placement of Workloads in Composable Data Center Networks

This paper studies the energy efficiency of composable datacentre (DC) infrastructures over network topologies. Using a mixed integer linear programming (MILP) model, we compare performance disaggregation at rack-scale and pod-scale selected electrical, optical hybrid topologies relative to traditional DC. Relative DC, results show that physical is sufficient for optimal when topology adopted resource components are allocated in suitable manner. The also enables DCs. logical DC servers an topology. rack-scale, server resources within each rack marginal fall total power consumption (TDPC) due improved demands placement. Hence, adaptable infrastructure can support both memory (access) latency sensitive insensitive workloads enabled. We conduct study adoption micro-service architecture Our increasing modularity improves DCs, but disproportionate utilization persists. A combination micro-services achieved up 23% reduction TDPC by enabling efficiencies. Finally, propose heuristic efficient placement DCs which replicates trends produced MILP model formulated this paper.