The Design And Implementation Of Monitoring And Warning System On Freelink Call Center Platform Based On Ganglia

With the development of computer technology, the performance of personal computer, server, workstation and high-speed network system continuously enhanced. Large computer systems have changed from traditional mainframes, super-computers to the cluster system consisting of multi-machines. And with the development of cluster technology, virtualization technology appeared to take further advantage of the performance of clusters. On the basis of the cluster's physical hardware, the virtualization technology transfer physical cluster resources into a virtual abstract cluster resources, which improved the utilization rate of cluster resources. Confronted with cluster system which consists of large-scale nodes and flexible configure, the traditional methods for monitoring mainframes, super-computers could not work anymore. In this case, it poses a severe challenge for monitoring and early warning the cluster system.Based on the platform of Free Link call center virtualization clusters and open-source cluster monitoring system of Ganglia, this paper designs and implements a monitoring and early warning systems. The System provides capabilities of easy-to-use monitoring and early warning to users, and helps users to better maintain the Free Link call center cluster system.The system is divided into three layers: system core business layer, data persistence layer, and web services layer. And each part is detailed design and implementation by this paper.The system core business layer is divided into two subsystems: the monitoring subsystem and warning subsystem. The implementation of monitoring subsystem is based on Ganglia, a open-source monitoring data collection tool and expands Ganglia's function of collecting data. By using Ganglia's python extensional module, the Free Link call center application server runtime data can be obtained. These data including extensional data will be saved in Mongo DB database, pushed to users and presented to the user with visual charts. The warning subsystem designs early warning service, which is based the data that the monitoring subsystem collected. The warning subsystem allow users to customize warning event trigger conditions according to their requirements. And the subsystem will compare the event trigger condition with monitoring data. The system will judge whether to trigger the event according the compare results. If the warning event happens, then a notification message will be sent to administrator.Data persistence layer encapsulates the data access operations, provides consistent data operate interfaces and masks complicated database connection and management operations. So that other businesses within the system can be employed in the form of an object-oriented operating data to facilitate database management and maintenance.The web services layer encapsulates system's core business and provides the monitoring and early warning service to a Web view, which is designed to provide users with a better use and operate visual interface.Finally, the routing system functional test and performance test are conducted. Functional test using black-box method, the test results show that the design of the system works well and achieves all the expected features. The performance testing by simulating a large number of concurrent requests on the system CPU, memory, and network load test. The test results show that the system designed in this paper has a large concurrency and high reliability, and can meet the needs of business needs of Free Link Call Center.