Premium
Evaluation of self‐organizing and self‐managing heterogeneous high performance computing clouds through discrete‐time simulation
Author(s) -
Giannoutakis Konstantinos M.,
FilelisPapadopoulos Christos K.,
Gravvanis George A.,
Tzovaras Dimitrios
Publication year - 2021
Publication title -
concurrency and computation: practice and experience
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.309
H-Index - 67
eISSN - 1532-0634
pISSN - 1532-0626
DOI - 10.1002/cpe.6326
Subject(s) - cloud computing , scalability , computer science , software deployment , distributed computing , resource allocation , resource management (computing) , resource (disambiguation) , database , computer network , software engineering , operating system
Recently, a new management and resource delivery model based on self‐organization and self‐management (SOSM) principles for heterogeneous cloud infrastructures was proposed. The novel architecture was designed and implemented for the provision of cloud resources for delivering high performance computing services. The cost of deployment of such framework on large‐scale cloud infrastructures is prohibitive, thus discrete‐time simulation techniques were adopted in order to evaluate its efficiency and scalability. In this work, extensive simulation experimentation is being reported and discussed, concerning four evaluation criteria, for comparing the traditional centralized resources allocation scheme to the alternative of SOSM. From the results obtained, it can be derived that SOSM can provide improved service delivery, computational efficiency, power consumption, scalability and management of resources for millions of cloud nodes, compared to the centralized resource allocation approach.