Research On QoS Of Real-time Multimedia Data's Transmission In Cognitive Network

Real-time Multimedia Data's Transmission has the features of strong real-time, large volume of data and burst strong. With the rapid development of real-time applications such as video conferencing and VoIP, delay-sensitive multimedia traffic will account for more than 80% of the bandwidth. The existing network QoS mechanisms are difficult to meet the next generation of network's quality of service requirements of various applications. Therefore, we must design a comprehensive architecture of QoS to make sure the real-time applications work with the non-real-time ones fairly, reasonably and effectively.The cognitive network, with self-management, self-learning and self-tuning capability, imports the artificial intelligence technology into the complex network management technology to achieve that the network can be controlled, managed, trusted. It is considered to be the future direction of network management technology. Currently, the researches on real-time multimedia transmission's QoS are a hot topic.This article studies two respects of the intermediate routing nodes and terminals. For the intermediate routing nodes, we propose a cognitive QoS method based on parameter sensitivity. The applications are classified as different combinations of parameters. The cognitive layer can calculate the probability of the packet drop dynamically. So it solved the problem that useless packets are sent first when the network congestion occurs and makes the forwarding strategy more reasonable. The actual performance of the network has also improved. For the terminals, we studied the binomial congestion control algorithm in depth and pointed out that the selection of value k and l is essentially the selection between the utilization of the data and the smoothness. It provides a range of optional space for the real-time multimedia applications to select the appropriate congestion control protocol.We also use the NS-2 simulation platform to simulate the applications of the two technologies to verify the correctness of the theories. Finally, we conclude the achievement and defects and point out the focus of future work.