The Research And Design On Resources Allocation Of Enterprise Private Cloud Computing Platform

Since enterprise IT system has been becoming more complex and more extended, wasting of servers’hardware is increasingly severe. A vast amount of practices reveal that almost every company has many application systems which utilize hardware resource uneven. The conservative method of hardware selection is configuring resources as the standard of peak; otherwise, users experience would be impacted at the heavy-load occasion. However, the utilization of those systems might be only10%at free time and the unoccupied resource cannot be used by other systems. Enterprise private cloud can be used as rapid responding method for companies’ hardware requirements of IT systems. Private cloud platform uses virtual machine as the basic element of resource, so that it can allocate resource at peak time and recycle resource at free time dynamically. By profound research of other papers and realistic of many companies, this paper proofs key points and difficulties. This paper describes the method of creating standard structure of private cloud.Revealed by researches, in order to create a stable platform and provide adequate capacity and expandability, some efficient method should be followed to calculate resource needs accurately. Based on those needs and specifics, this paper mainly describes the mechanism of dynamic resource allocation, selection of performance counters and baseline. This paper also discusses about methodology of how to measure accurate baseline and adjust hardware resource by adding or removing virtual machine pool. Also an experimental model is designed to describe and proof the proof feasibility of dynamic resource allocation.This paper contains a vast amount of experiment’s data, testing result and comparing information of international research or solutions of private cloud. By doing this, this paper creates a set of pre-requisites and building methods of creating enterprise private cloud which meets exact needs of companies. By adding application specified performance indicators, complete error handling mechanism and edge detection mechanism to extend this model, it can be used with more advanced functions of dynamic resource allocation.