Authors:
Abeer Al-Fadhel
;
Paulvanna N. Marimuthu
and
Sami J. Habib
Affiliation:
Kuwait University, Kuwait
Keyword(s):
Utilization, Server Consolidation, Redesign, Optimization, Simulated Annealing.
Related
Ontology
Subjects/Areas/Topics:
Agents
;
Applications
;
Artificial Intelligence
;
Bioinformatics
;
Biomedical Engineering
;
Enterprise Information Systems
;
Information Systems Analysis and Specification
;
Methodologies and Technologies
;
Network Optimization
;
Operational Research
;
Pattern Recognition
;
Resource Allocation
;
Simulation
;
Software Engineering
Abstract:
In this paper, we have explored the utilization of existing servers within an enterprise information network (EIN), and we have proposed redesign operations on servers to identify and remove the low-utilized servers. The low-utilized servers consume unnecessary power and increase the operational and maintenance cost. The removal of low-utilized server is viewed as an EIN redesign problem, which removes the low-utilized servers within the EIN and re-distributes the clients of the purged servers to the remaining servers, thereby reducing a portion of expenditure on maintenance and operation. We have proposed three approaches on distributing the clients of removed servers and the approaches are; single server pure random distribution, selective distribution and multiple servers pure random distribution. We have employed Simulated Annealing to search for best possible random server/servers in order to distribute the workload of the removed server, thereby improving the utilization of the
remaining servers. The simulation results for a given EIN with 10 servers and 25 clusters show that our proposed server consolidation approaches improve the initial average server utilization of around 25% to 60%, 68.5%, and 90% respectively in the proposed three methods.
(More)