A Load Balancing Model Based On Traffic Prediction In Cognitive Networks

Existing in the traditional network resources utilization rate is low, problem solving localised and business QoS cannot assure, etc. Along with the rapid development of Internet, people on the network's QoS raised more requirement, cognitive network points to solve the problem of above to provide the possibility. Cognitive network, on the basis of traditional network, join learning reasoning, intelligent decision to such mechanism. Compared with the traditional network, cognitive network have more intelligent, and can get the network status of perception, according to the actual operation of the entire network make more sensible decisions, provide more high-quality network QoS, improve the network resource utilization.This paper firstly do in-depth study on cognitive network, traffic prediction and load balancing the domestic and foreign research situation and development trend of the careful analysis. This paper combine cognitive network, traffic prediction and balanced load, reasoning and intelligent decision concept such as network traffic prediction and load balance to realize cognitive network, making the network has the more intelligent and rationality.Based on the introduction of the traditional network traffic prediction and load balance by use of the method was discussed based on both applied in cognitive function and significance of network, and put forward cognitive network traffic prediction model based on two states markov chain's, based on traffic prediction of load balance control model. At last, this paper describe the cognitive network traffic prediction and load balancing subsystem models in the specific algorithm, the simulation results, the design and realization.In the end, the paper builds an experimental platform, this paper introduces the cognitive network traffic prediction and load balancing subsystem design development and test process, the main job content, including system deployment, system design, system function is introduced, the system module detailed design and system test, etc.