Approach To Resource Monitoring And Evaluation In Situation Of Cloud Computing Platform

Nowadays, the age of big data has come and gradually had a profound influence on our lives. The development of cloud computing attracted much attention from different domains and the concept of ‘cloud’ had deep effects on the technology improvement by the way of exploring and practicing about the cloud computing. Google has published papers to give a detailed explanation about cloud computing in 2003. Amazon commercialize the cloud computing in 2006. Recently, ‘Cloud’ has stepped into our life as a basic service. Some computing platform has been forced to appear under this environment and Open Stack is the most representative one.Firstly, this paper elaborates the related concepts about cloud computing, cloud platform and virtualization and introduces some popular platforms and their features. Secondly, the paper compares and analyzes some applications of monitoring technology in different scenarios. Then, the paper gives a detailed explanation of situation assessment, features and assessment factors. Meanwhile, we compare the existing situation assessment methods. Above all, we design and implement single send/single receive multi broadcast topology structure monitoring system deployment proposal based on Open Stack cloud platform. And then we propose a cloud platform situation assessment method by the way of dealing with monitoring data.This paper mainly improve the default multi send/multi receive multi broadcasting topology structure to prevent each node to receive the index data from all the other nodes in order to make full use of the CPU resource and reduce the running spending large-scale clusters. What’s more, gmetad protecting process of each cluster use RRD file to aggregate resource data information and collect them to master nodes. We use UI modules developed by some advanced technologies such as Bootstrap, Highcharts, node.js to deploy main master nodes. On one hand, this can provide the monitoring system on realtime changing data for the administrators. On the other hand, this helps provide the entrance to tasks submission and result examine to the users.The main goal of the research on situation assessment methods in this paper is to mine the potential value of data in the platform. The situation assessment methods utilized in the paper mainly refer to the D-S evidence inference and decision judgement. Themethods use monitoring data of ‘dot’ and ‘segment’ resources and log information to assess and predict the health and running state of the platform. Meanwhile, our methods could reflect the own features of the tasks by the ways of changing trends of monitoring data. We also use the tasks completion states to judge the stability of the whole platform or some clusters and utilize fault tips to analyze the root cause of faults and the situation the platform. We adopt this approach to guarantee solving the problems of the platform in time.Cloud monitoring system is necessary for a reliable cloud platform and it provides support for platform availability; Situation assessment methods have direct effects on maximizing the data value of cloud platform. Finally, this paper takes the function tests and performance tests of the implemented system. The effectiveness, correctness and stability has reach the goals of the paper.