Research And Implementation Of High Reliability And High Performance Test Data Center System

With the continuous improvement of technology, there is a new peak in China's navigation. There will be a lot of experiments and data in the process of developing the navigation. The management of the experiments and data processing will be the core of the future development. What is described in this article is the core component of a subsystem. This system needs to be able to guarantee the continuity of the business and adapt to the continuous high load operation. At the same time, it can receive, storage, forward, manage and use the test data. Additionally it can provide unified management as a core with service.In order to meet the requirements for function, performance and reliability of the system, this article describe the test data center system based on overall system architecture, reliability design and high performance design.About overall architecture, we divided the data center system into network layer, core service layer, storage layer and operation and management layer. It described the design of each layer, the construction scheme and system architecture combined with the high reliability and high performance requirements.About reliability design, we carry on high reliability design from four aspects combined with business requirements. In network layer, we mainly access vlans and link redundancy technology to improve its reliability. In terms of business software, we mainly adopted the design of error avoiding, error detection, fault-tolerant and error correction to improve reliability. We bought the X3850 server which is in high performance and used the server virtualization technology. Through the VMware vsphere software, we greatly reduced the number of servers and improved the reliability of the system. Moreover, we improved the utilization rate of resources and reduced the construction cost. To improve the storage system, we purchased high-end disk array without a single point of failure; used redundancy link for all component interconnection; adopted different level of RAID storage, and made a data backup to ensure the reliability of the storage system. Meanwhile, made a lot of research and practice for multi-optional technology nodes to select the best technical solution which faced by this system. At last, we make a reliability test via both hardware and software sides. It showed that the test results meet the requirements of previous analysis and proved the feasibility and high reliability of the system.About high performance design, this article starts with the perspective of core service layer and storage layer. In terms of core service layer, we improved its performance through software architecture optimization. In the process of the software design, we used multi-threading to improve utilization ratio of resource, used good design patterns and focused on garbage collection to ensure the performance of the software. In storage layer, we joined the cache server in order to improve the read/write performance. Meanwhile, we optimized the disk type selection, database selection and RAID design to ensure the high performance of storage layer. The test showed that the various features completely accord with the standard of all performance indicators, and proved the high performance of test data center system.Test data center system can unify storage planning, which effectively protect and integrate the data, and eliminate information islands. Besides, the system can fast adapt to the needs of the business via layered architecture, It not only provide the unified management with service as the core, but also realize the late convenient extension. Besides above all, the test data center system can also quickly finds valuable information from large amounts of data. This system solved the problem of data storage and management and remedied the defects of the poor in time effectiveness and the unreliability of data.