A QoS and profit aware cloud confederation model for IaaS service providers

  • Authors:
  • Amit Kumar Das;Tamal Adhikary;Md. Abdur Razzaque;Eung Jun Cho;Choong Seon Hong

  • Affiliations:
  • University of Dhaka, Dhaka, Bangladesh;University of Dhaka, Dhaka, Bangladesh;University of Dhaka, Dhaka, Bangladesh;Kyung Hee University Global Campus, Deokyoungdaero, Giheung-gu, Yongin-si, Gyeonggi-do, South Korea;Kyung Hee University Global Campus, Deokyoungdaero, Giheung-gu, Yongin-si, Gyeonggi-do, South Korea

  • Venue:
  • Proceedings of the 8th International Conference on Ubiquitous Information Management and Communication
  • Year:
  • 2014

Quantified Score

Hi-index 0.00

Visualization

Abstract

Cloud confederation is the union of cloud data centers that allows outsourcing user requests by a cloud provider during peak hours to liberate it from the constraint of limited physical resources. Federation extends some of the features of cloud like low cost, scalability, robustness and availability while increases the price and revenue of the providers along with maintaining a high QoS by effectively utilizing the resources of the data centers. Existing works on cloud federation do not consider all three categories of instances and hence cannot maximize their profits. Moreover, the process of establishing cloud federation and taking cost efficient decisions based on QoS parameters has not yet been provided in recent works. In this paper, we have designed a data communication model for the members of a cloud confederation, which exploits clear understanding of the usage pattern, types of requests in addition to infrastructure expenses. We develop an algorithm and a model for cost calculation, which enhances the decision making process over all the VM types (on-demand, reserved, spot) to increase resources utilization and profit. The simulation results, conducted on CloudSim, indicate that, our proposed model and algorithm enhances the profit, utilization, and QoS in a cloud confederation environment compared to a number of state-of-the-art approaches.