Research And Implementation Of QoS Class Filter Mechanism Based On RSVP

As a tremendous growth of multimedia technology applying in IP network, network should have the ability to recognize and distinguish different types of communication, and provide corresponding service. QoS technology intends to solve the problem. Furthermore, RSVP is utilized to transfer different requirements of QoS and set up end-to-end QoS. The biggest obstacle to QoS implement is network congestion. Reasonable strategy is discussed in this thesis to discard the dross and select the essential to reduce the effect of the network congestion. Regardless of priority service, removing unimportant data will cost less decoding quality than the lost of essential data caused by network congestion. Therefore, this thesis will do research on the method of removing nonessential data during video stream transmission so as to reduce the chance of network congestion and guarantee the end-to-end QoS. The compressed encoding audio and video data stream is becoming the principal part of the internet data transmission. And the real time video transmission needs the most bandwidth resources, which is the issue the thesis focus on. This thesis firstly induces the characteristics of compressed encoding video stream by analyzing video encoding technology, and writes programme implementing encoding & decoding function of MPEG-4; Secondly proposes a QoS Class filter Mechanism Based on RSVP, and expatiates its principle, feature and application; finally validates the effect and application area through simulation experiment on a designed prototype. In the QoS class filter mechanism, packets in the same session are classed by their importance, and RSVP signaling is extended to define corresponding classes of service. RSVP is used as signaling protocol and QoS filters form the active components in the intermediate nodes. They remove unnecessary information from continuous video streams to protect the transmission of essential ingredients, and QoS can be scaled down to the individual requirement of subgroups or single members within a heterogeneous group communication. It is proved that this mechanism can solve the network congestion problem effectively and improve the end-to-end QoS when bandwidth is a scarce resource. The innovation of this thesis lies in the following points: Using the characteristics of compressed encoding video stream to class different packets in the same session based on their importance; Extending RSVP signaling to transmit QoS class information; Simple QoS filter mechanism is set up in the intermediate nodes to remove unnecessary information; Utilizing Error-Correcting function of MPEG-4 to reduce the impact to the video decoding quality caused by the lack of unnecessary information.