Optimization analysis of management operation for a server farm. Issue 3 (3rd May 2020)

• Record Type: Journal Article
• Authors: Chang, Fu-Min
• Is Part Of: Quality technology & quantitative management. Volume 17:Issue 3(2020)
• Page Start: 307
• Page End: 318
• Publication Date: 2020-05-03
• Subjects: Server farm -- matrix-geometric -- cost model -- optimization -- unreliable; Quality control -- Periodicals; Quality control -- Statistical methods -- Periodicals; Industrial management -- Periodicals; Industrial management; Management -- Research -- Methodology -- Periodicals; Qualitative research -- Periodicals; Management; Quality control; Quality control -- Statistical methods; Periodicals; 658.00721
• DOI: 10.1080/16843703.2019.1618535
• Languages: English
• ISSNs: 1684-3703

Abstract

ABSTRACT: This paper analyses a server farm with a simple management operation, which is desirable to reduce power consumption. A block of servers is named as 'reserves'. Depending on the number of jobs staying in the system, power-up and power-down thresholds are employed to control the state of the reserves. In addition, the process of power-up is not immediately. During the period, the servers cannot serve jobs but still consuming power. The server may be subject to be breakdown. This system was modelled by an infinite capacity queueing system and analysed by matrix-geometric technique. The matrix-geometric method is applied to compute the rate matrix and the stationary probabilities. A cost model is formulated to search the optimum number of permanent server and the optimum power-down threshold. The optimization tasks are carried out by the direct search method. Both analytic processes and numerical results provide very useful and helpful information for decision-makers.