Aura: A Type-Safe, Multi-Paradigm Cloud Simulation Framework in Scala 3

Abstract

Cloud applications today run across virtual machines, serverless functions, container orchestrators, edge nodes, and federated tiers that share the same physical hosts. Existing cloud simulators model these paradigms one at a time. We present Aura, a discrete-event simulation framework written in Scala 3 that models all five within a single actor-based engine. We use Aura to measure an effect that no prior simulator can: when serverless cold-start bursts overlap with Kubernetes pod-binding waves on shared hosts, aggregate energy rises 9.3% above the sum of the same workloads run in isolation, and peak per-host CPU climbs from 74% to 93%. On a workload-density scaling benchmark with fixed infrastructure, Aura runs 50,000-workload simulations 11× faster than CloudSim Plus; CloudSim Plus is faster on small scenarios where its lower per-entity overhead dominates. Serverless cold-start behavior matches published AWS Lambda measurements and the Azure Functions production trace. Aura is released as open-source software.