FFMRA: A Fully Fair Multi-Resource Allocation Algorithm in Cloud Environments

Authors: Hamzeh, H., Meacham, S., Khan, K., Phalp, K. and Stefanidis, A.


Start date: 19 August 2019

Pages: 279-286

Publisher: IEEE

The need for effective and fair resource allocation in cloud computing has been identified in the literature and in industrial contexts for some time now. Cloud computing, as a promising technology, offers usage-based payment, ondemand computing resources. However, in the recent decade, the growing complexity of the IT world resulted in making Quality of Service (QoS) in the cloud a challenging subject and an NP-hard problem. Specifically, fair allocation of resources in the cloud is one of the most important aspects of QoS that becomes more interesting especially when many users submit their tasks and requests include multiple resources. Research in this area has been considered since 2012 by introducing Dominant Resource Fairness (DRF) algorithm as an initial attempt to solve the resource fair allocation problem in the cloud. Although DRF has some good features in terms of fairness, it has been proven inefficient in some conditions.

Remarkably, DRF and other works in its extension are not proven intuitively fair after all. These implementations have been unable to utilize all the resources in the system and more specifically, they leave the system in an imbalanced situation with respect to each specific resource. To tackle those problems, in this paper we propose a novel algorithm namely FFMRA inspired by DRF which allocate resources in a fully fair way considering both dominant and non-dominant shares.

The results from the experiments show that our proposed method provides approximately 100% utilization of resources and distributes them fairly among the users and meets good fairness properties.

The data on this page was last updated at 05:09 on February 27, 2020.