Towards a General Framework for Studying Resource Management in Large Scale Distributed Systems

P. 79–96.

Oleg Sukhoroslov, Maxim Gorokhovskii, Roman Ilgovskiy, Tagir Kuskarov, Semenov Y., Andrey Vetrov

Resource management is a critical part of modern large-scale distributed systems and an active area of research. The diversity of systems, workloads and requirements have lead to many domain-specific problem statements, solutions and simulation tools proposed by researchers, which are hard to compare and reuse. In this paper, an overview of related resource management problems is made and a general simulation framework to support studying such problems is presented. The framework has modular architecture, and is built around the generic simulation core and a set of basic resource models, which can serve as a common foundation for development of domain-specific models, components and simulators. The framework prototype has been evaluated and proved to be flexible and adaptable to different problem domains, while achieving performance and scalability needed to support the modeling of modern large scale problems.

Keywords: scheduling, simulation, resource management, cloud computing, distributed systems

In book

Материалы IV Международного семинара по информационным, вычислительным и управляющим системам для распределенных сред (ICCS-DE 2022)

Иркутск: ИДСТУ СО РАН, 2022.