Analysis And Design Of Business-Driven Capacity Planning Model In The Cloud

Traditionally, the enterprise built its IT infrastructure in advance based on its peak or average demand. Thus the enterprise took the cost of construction, operation and maintenance of IT infrastructures. With the emergence of the cloud computing paradigm and the continuous search to reduce the cost of running Information Technology (IT) infrastructure, we are currently experiencing an important change in the way these infrastructures are assembled, configured and managed. Cloud providers have taken the burden of the cost of IT infrastructures from the enterprise, so it is left to the enterprise to make an optimization provisioning plan according to computing resources provided by cloud providers.The issue that we address in this paper is how cloud users could make a provisioning plan for computing resources. This work is mainly the following aspects:1) Introduce the background knowledge about Amazon EC2, mainly introducing the concepts of region, AMI, Instance. Then introduce three purchasing options used by current popular cloud platforms, i.e. on-demand purchasing option, reserved purchasing option and spot purchasing option. And make a statistical analysis on spot instance historical prices to find out the way Amazon prices the spot instance.2) Make a demand analysis of capacity planning, and propose the design theory about long-term capacity planning and short-term capacity planning. Analyze the roles and the targets of long-term capacity planning and short-term capacity planning separately.3) Propose a long-term decision model for provisioning computing resources. Make a detailed introduction of design concept and goal achieved. Compare the model with other four capacity planning methods to prove the superiority of the proposed model.4) Propose a short-term decision model and an algorithm based on this model. Introduce the basic idea of the model, i.e. Branch and Bound method. Carry out a simulation experiment to evaluate the efficacy of the algorithm.5) Propose a high-level architecture for a short-term automation capacity management system.In this paper, we propose a systematic set of capacity planning methods. Long-term capacity planning can determine a cost-optimal resource provisioning schedule that satisfies the demand, while optimizing the cost. Short-term capacity planning help users to solve the unpredictable application demand problem to assist the long-term capacity planning. The experimental results show these capacity planning models have advantages on cost or time.