基于云平台的大规模工业园区资源集约化控制方法

Resource Intensive Control Method of Large-scale Industrial Park Based on Cloud Platform

为了解决传统分布式控制部署环境繁琐,令现阶段资源控制变得更加复杂,容易产生服务延迟和稳定性低的问题。通过云平台研究了一种大规模工业园区资源集约化控制方法。以服务思想为核心建立大规模工业园区资源集约化控制框架,为用户提供统一的使用平台。设计云平台由云控制节点控制,按照需求对存储节点进行扩充。框架中的功能层为大规模工业园区资源集约化控制的重要部分,包括性能监测、资源调度、自适应控制与报警四个部分。虚拟机放置遍历全部可能的云平台状态,将状态变化代价最小云平台状态当成虚拟机控制策略。基于成本的最低消耗采取资源调度策略,组建动态任务划分模型。结合总体考虑节点量、输入数据类型、进行调度时所耗时间以及资源回收所需时间等影响因素做出深入分析。结果表明:所提方法控制下网络拥塞率符合要求;服务延时低、稳定性高。可见所提方法整体性能优。

In order to solve the problem that the traditional distributed control deployment environment is cumbersome,which makes resource control more complex at the present stage,and prone to service delay and low stability,a resource intensive control method for large-scale industrial parks is studied through cloud platform.With the idea of service as the core,a large-scale industrial park resource intensive control framework was established to provide users with a unified use platform.The design cloud platform was controlled by the cloud control node,and the storage nodes were expanded according to the requirements.The function layer in the framework was an important part of resource intensive control in large-scale industrial parks,including performance monitoring,resource scheduling,adaptive control and alarm.The virtual machine was placed to traverse all possible cloud platform states,and the cloud platform state with the lowest cost of state change was obtained as the virtual machine control strategy.Based on the strategy of resource scheduling with minimum cost consumption,a dynamic task allocation model was established.The factors such as the number of nodes,the number of sub-tasks,the types of input data,the cost of scheduling time and the cost of recovering resource time were comprehensively analyzed.The results show that the network congestion rate meets the requirements under the control of the proposed method,and the service delay is low and the stability is high.It can be seen that the overall performance of the proposed method is excellent.