如何对资源进行合理有效的分配一直是云计算领域的热点问题.由于传统的云资源分配算法不能有效屏蔽底层硬件资源的异构化以及不同层的云服务类别,本文提出一种多参数资源打包的方法,构建出基于离散事件系统的云资源分配控制模型,并给出了合理的资源分配算法.算法通过计算服务器端各资源包容量参数与客户端资源需求量参数的贴近度来定义资源分配事件,并对事件发生与否的状态实行分层控制,最终使得整个资源分配系统到达可接受状态.实验表明,基于离散事件系统的云资源分配模型,能够保证在可接受状态下,不仅每个用户的资源请求能够得到合理的分配,且能实现云资源利用率最大化.
How to allocate resources reasonably and effectively has been an important problem in the field of cloud computing. For traditional cloud resource allocation algorithms can not effectively shield the isomerization of underlying hardware resources, and they also can not neglect the categories of different layers' cloud service, this paper proposes a multi-parameter resource packing method, and constructs a cloud resource allocation control model based on discrete event system. Besides, it gives a reasonable resource allocation algorithm. By calculating the close degree of capacity parameters from each resource package over the servers and the resource demands of the clients, the algorithm defines the resource allo- cation events, and hierarchically controls the states' occurrence or nonoccurrence. Finally, the entire re- source allocation system will reach an acceptable state. The experiments show that the cloud resource allocation model based on the discrete event system can guarantee that under the acceptable state, not only each resource request can get a reasonable allocation, but also the resource can get a maximal utilization.