Design And Realization Of Service Context Subsystem In Context Aware Autonomic Network

With the continuous emergence of new technologies and applications, the traditional Internet has achieved great success. However, with the increase of application types and technologies, the traditional network is facing enormous pressure. On the one hand, the data transmitted in the internet and their communication methods are now diversiform, not only including the traditional text data transfer and the point to point interaction; On the other hand, the shortcomings of the traditional IP network deployment techniques in hierarchical network design, quality of service, network security, and other properties are increasingly exposed. Therefore, it's urgent to make the necessary improvements to enhance the ability of Internet service.According to the needs of network communication in the future and the failure of the traditional internet, "Research of Pervasive context aware autonomic network architecture model"(CAANM) studied a new type of network architecture and achieved the autonomic computing/ communication techniques to build a pervasive context aware autonomic network model which will support the Self-management, heterogeneous communications and QoS (Quality of service).Through the establishment of CAANM, the feasibility of the autonomic network architecture was verified.The research issue in this thesis "Design and Realization of Service Context Subsystem in Context Aware Autonomic Network" is the sub-issue of CAANM. This topic focuses on the deficiency of current network traffic management which only by the basic flow control and simple and rough source address filtering, but lack the aware capacity of users and application, and then put forward a service context subsystem which will achieve the deep service context awareness intelligently and proactively in the management of network traffic. Reference to the service context aware technologies and the function requirements of CAANM, all the modules of service context subsystem have been designed and realized. They were inosculated as a whole to provide supports for the autonomic behaviors of other modules in CAANM, which will helpful for the autonomic QoS management.For prototype realization of the subsystem, a real time video stream key frame identify algorithm for QoS named RVKI is proposed, considering that it is popular and requires a higher QoS. In this algorithm, the frame type of IP packets could be identified at the real time multimedia server and be marked. This result can be used for some autonomic QoS management to improve the success of IP packets transmission to the client, which carrying real time multimedia key frame information, when network congestion occurs but could not decrease the packet loss of this application. Test has shown that RVKI can effectively dissect key frame in the real time multi-media stream and helpful for QoS.At last, the major work of this thesis is summarized, and the possible further work and research directions improvements are pointed out.