Green and FairWorkload Distribution in Geographically Distributed Datacenters

Abstract

To accommodate user applications and requests, service providers host large computer systems over an infrastructure of distributed datacenters in different geographical locations. Consequently, energy and power consumption increases, increasing electrical and cooling costs. Hence more attention is paid to green computing and exploiting renewable energy sources, along with decreasing electrical costs so that these datacenters are operated efficiently and effectively, with less impact on the environment. In this paper, a socially responsible scheduling scheme that distributes load from several regions to geographically distributed datacenters will be extended. Optimal electricity, social and latency costs will be sought within service level agreements, ensuring fairness between different regions and highlighting its importance. Adopting standard gradient methods and Lagrange multipliers, an optimization algorithm called Green-Fair is proposed that provides optimal selection of datacenters. Experimental results show the effectiveness of the Green-Fair algorithm in utilizing green energy sources, reducing the number of SLA violations significantly, reducing latency costs and preserving fairness requirements by the SLAs while having little effect on the service providers' overall operational costs.

Data Center Architectures: Challenges and Opportunities

“…Power management could be achieved by distributing workload across data centers efficiently. An optimized algorithm called Green-Fair has been proposed by Alawnah et al [45] for workload distribution in geographically distributed data centers based on the standard gradient method. Results show that Green-Fair distributes the work between different regions in such a way that no region will be idle.…”

Moment-Generating Algorithm for Response Time in Processor Sharing Queueing Systems

“…For example, users wait, on average, just over nine seconds for a web page to load [27] before opting for more reliable performance from competitors. The same argument applies to delays in data centers [14,16] as part of quality of service (QoS) standards, which is incorporated, along with operational and energy costs [15,18], into service level agreements (SLAs). Whether it's using smartphones to download files using WiFi or streaming web content on the cloud, the delay principle still applies.…”