The Research On The Construction Technology Of Large-Scale Experimental Environment Based On Virtualization

In today’s IT technology field, no matter software, hardware, systems or complete solutions, it is essential to build an emulated experimental environment to meet the needs of their tasks, such as technical experiments at the beginning of new product research and development, product test at the end of the design and development, and technical training for the specific product, etc. However, it is not a good idea to directly construct a large-scale networking information system by real-world equipment due to several drawbacks. Firstly, it is costly in devices, human labor and maintenance. Secondly, the construction and recovery of the physical environment, and scheduling between tasks are very time consuming. The resources in the physical environment are isolated, un-extendable, and can’t share with others. Thirdly, it is hard to do experiments under the working environment. Meanwhile, the experiments can’t fully be developed under a R&D environment because of the absence of feasible conditions. To solve the above problems, this thesis developes a large-scale experimental environment construction approach based on virtualization.The related experimental environment construction techniques in the literature are reviewed, e.g. Live, Constructive Simulation and Virtualization approaches. By summarizing the pros and cons of the current techniques, a novel approach is proposed to meet the need of our project. The contribution is as follows:(1) The "six features" requirements of systems and ability requirements of experimental environment construction are analyzed and summarized. Taking them as the input, "three vertical and two horizontal" function structure model and "four step" experimental environment construction process framework based on experimental resources are proposed. They build up the foundation to realize the technology and system design.(2) An approach to describe the experimental environment requirements based on the essential characteristics is proposed. The method establishes the specifications for experimental resource classification and description. Moreover, the elements comprised of the target experimental environment and essential characteristics at communication networks, computing facilities and application system levels are refined through the proposed method. A description model of experimental environment requirements is established to provide the guidance for the description and generation of target experimental environment. An experimental environment construction method based on LVC is also proposed. It focuses on the construction technology of computing and communication nodes of the target experimental environment based on virtualization, and the experimental resource mapping technology based on topology segmentation. They provide technical support for the experimental environment construction system.(3) The general framework, physical network and system organization are designed. It focuses on resource registration management, experimental environment configuration and deployment. It introduces the subsystem functions, deployment model, hierarchical implementation architecture and the main function implementation processes, as well as the main interfaces, to meet the needs of experimental environment construction for test task team.(4) Based on the hardware and software environment of the experimental environment construction system prototype, with focus on the test requirements of the replication of a typical networked information system, the deployment performance, the essential characteristics of communication networks and computing facilities are tested. The test results demonstrate the usability and effectiveness of the construction technology and system of experimental environment based on virtualization. They can meet the usual requirements of test tasks in function, performance, security, etc.The techniques in this study can partially solve the problems occurred in the product development or promotion, in which a large number of experiments can’t be conducted because of the lack of suitable environments. It can provide a private cloud environment at the beginning of project debugging, functional test, and new technique experiments. It can meet the requirements of various test tasks, and are of great value in engineering applications.