SLA-based optimisation of virtualised resource for multi-tier web applications in cloud data centres

Abstract

Dynamic virtualised resource allocation is the key to quality of service assurance for multi-tier web application services in cloud data centre. In this paper, we develop a self-management architecture of cloud data centres with virtualisation mechanism for multi-tier web application services. Based on this architecture, we establish a flexible hybrid queueing model to determine the amount of virtual machines for each tier of virtualised application service environments. Besides, we propose a non-linear constrained optimisation problem with restrictions defined in service level agreement. Furthermore, we develop a heuristic mixed optimisation algorithm to maximise the profit of cloud infrastructure providers, and to meet performance requirements from different clients as well. Finally, we compare the effectiveness of our dynamic allocation strategy with two other allocation strategies. The simulation results show that the proposed resource allocation method is efficient in improving the overall performance and reducing the resource energy cost.

Keywords:

• performance optimisation
• dynamic resource allocation
• virtualisation
• cloud data centre
• service level agreement (SLA)
• multi-tier web application

Acknowledgements

This work is supported by the National Natural Science Foundation of China (No. 61174169 and No. 61033005), the National High Technology Research and Development Program of China (863 Program, No. 2012AA040915) and the National Key Technology Research and Development Program (No. 2012BAF15G00).