Cloud computing simulation tools: Comparative analysis and the distributed execution gap

Abstract

Cloud computing simulation is the dominant methodology for evaluating resource provisioning policies, scheduling algorithms, and Quality of Service mechanisms at scale without incurring real infrastructure costs. Despite the maturity of the field, two practical gaps persist. First, the tool landscape has grown to over 33 distinct simulators spanning Infrastructure-as-a-Service, Fog, Edge, SDN, and energy domains, yet no consolidated guidance exists to help graduate researchers select the right tool for a given research question. Second, 31 of the 33 surveyed tools are constrained to single-threaded execution, a structural ceiling that forces researchers to conduct scaled-down experiments and extrapolate results to production-scale behaviour. This paper addresses both gaps. It presents a tool selection decision framework mapping 11 research scenarios to the most suitable simulators, derived from a comparative analysis of 33 tools across 7 key criteria. It further diagnoses the root causes of the single-threaded constraint and proposes a conceptual parallel simulator architecture — partitioned simulation spaces with conservative synchronisation — as a direction for future work. The authors conclude that while the ecosystem is rich and open-source adoption is high, the community requires both better selection guidance and a new generation of parallelism-aware simulation infrastructure to advance cloud computing research credibly at data-centre scale.