Performance Evaluation And Storage Optimization Strategy Of Gas Monitoring Big Data Platform

With the rapid development of Internet of Things technology,not only the Internet of Everything gradually becomes the future trendency of our society,but also the intelligent gas meters with the property to access the network are applied extensively.However,in front of the enormous market prospect of intelligent gas meters,the big data platform utilized to manage gas monitoring devices obviously faces tremendous challenges.The gas monitoring platform exists the following problems: 1)The traditional testing approaches of web application cannot accurately obtain the number of devices which link to the gas monitoring platform;2)The performance bottleneck of the current monitoring platform is difficult to locate effectively.In order to address the problems above,this thesis performs extensive investigation on the evaluation method of the gas monitoring big data platform and the optimization technology for the corresponding performance bottleneck.The main work and contributions of this thesis are made as follows:(1)An evaluation approach for the gas monitoring platform on the basis of simulator.Compared with the conventional evaluation approaches for the interface,this approach connects to the gas monitoring platform via the generation of the simulator,simulating the process of the connection and information interaction between the intelligent gas meter and the gas monitoring platform.Then collects the corresponding performance indicators in this process.Furthermore,the performance indicators are analyzed to obtain the maximum number of device and the performance bottlenecks of the gas monitoring platform.After the evaluation approach utilizing on the gas monitoring platform of Z Safety Engineering Company,we found the maximum number of device in access determined to be 3456,and located the performance bottleneck of the platform in the storage part.(2)A priority-based gas information storage strategy.Due to the characteristics of data that the abnormal information is more important than normal information in gas detection and the occurrence probability of abnormal information is much lower than that of normal information,we assigned different priorities to various types of gas information,and applied Redis to realize the caching of data.The low-priority gas information would firstly cache in the server,and then persist to the database in batches,while high-priority gas information would directly be stored in the database to ensure the real-time property of these data.Moreover,the threshold of cached data was verified by formula to obtain the optimal threshold which could access the most devices.The application of this storage strategy and the optimal threshold made the maximum number of devices connected to the gas monitoring platform increase to 78,000,which was 21-fold higher than before optimization,meanwhile strategy with optimal threshold got 13% more devices than other thresholds at least.Compared with other storage optimization strategies,it is also the best in terms of comprehensive ability.(3)Performance evaluation platform of gas monitoring platform based on the evaluation method proposed in this thesis is designed and acquired successfully.This performance evaluation platform allows users to freely set up and generate simulators,which could execute corresponding test plans and provide convenience for users to test the gas monitoring platform.It is also an excellent tool for users to apply the evaluation method proposed in this thesis,which could effectively verify the effectiveness of the approaches and the strategies proposed above.