Research And Construction Of Industry Cloud Computing Platform Based On Openstack

With the development of modern information technology, information technology to promote the rapid development of various industries, the demand for computing services is also gradually increasing. At the same time, rapid development of field bus technology in the industrial field, making industrial information for the calculation of the service demand also is increasing, to build "industrial cloud platform" has become the inevitable process of enterprise further to the intelligent. Although cloud computing technology has been used in many large-scale computing center, but the requirements of the industrial field for data transmission, different equipment performance, safety and reliability requirements are different from the normal cloud platform. So, how to construct a safe, reliable, low power consumption and high availability industrial cloud platform, is still a hot problem that many researchers have at present. To this end, the topic for the construction of industrial cloud platform to do the following research:(1) Aiming at the problem of server load performance prediction, this paper proposes a server load forecasting algorithm based on time series. For the realization of server load performance monitoring, this topic built a open-source cloud computing platform based on OpenStack and developed a resource monitoring module, finally realizing the automatic operation of the resource scheduling module. What's more, using the sliding autoregressive model averaging complete the server load index prediction, and collecting the real server data to experiment, this topic verified the effectiveness of the method proposed in this paper, for subsequent dynamic virtual machine migration provides a basis.(2) Aiming at the virtual machine scheduling strategy, this paper proposes a dynamic scheduling strategy based on energy consumption. In this paper, a cosine similarity of dynamic scheduling strategy based on, the cosine similarity to establish to server load for measure and the corresponding dynamic scheduling scheme knowledge base. And a business module has been designed and realized base on OpenStack. Using simulation data to verify this subject based on the effectiveness of energy consumption of dynamic scheduling strategy.(3) According to the platform deployment problem, this paper designs a high availability cloud platform architecture. In the test platform the resource scheduling module proposed in this topic was deployed. In order to eliminate the single point of failure. In the system of overall service, MySQL service and message queue service of redundant architecture design, to realize the high availability of cloud platform deployment. Finally, the real deployment environment to verify the availability of cloud platform.The research results show that our industrial cloud platform based on OpenStack, combined with the server load prediction algorithm based on time series to realize VM dynamic migration, which was reasonable on scheduling server resources. The dynamic scheduling strategy based on energy consumption and cosine similarity computing to build, can effectively reduce energy consumption of equipment. Through the cloud platform design for high availability deployments to verify availability, this topic provides the new train of thought for the industrial cloud platform construction.