Research On Qos Routing In Service Classification System

With the rapid development of network techonology, the Internet today has been evolved as a hybrid transmission network that supports multiple services such as data, voice, and image, as well as many other services. Compared to traditional network services, real-time services pose more critical requirements on network performance issues in terms of transmission delay, jitter, packet loss rate, etc. QoS-related techniques are capable of maintaining the overall quality of service for the networks, providing differentiated QoS for divergent services, and thus improving the utilization of network resources and satisfying the requirements from critical services. Based on thorough investigations in various problems and challenges among existing QoS routing techniques, we propose a novel QoS Routing model based on Service Classification, and the most significant advantage lies in its capability of supporting divergent quality of service for different network services. As the general framework, the model consists of several major functional modules including service identification/classification, routing, and statistical/analytical module. With respect to the network services that holds significant characteristics of a certain service category, we have designed the routing algorithms named Particle Swarm Optimization based on Fitness Evaluation Function(PSOFEF). This Master thesis describes the algorithm in details with the mathematical model and designs a novel two-layer-indexed routing table for supporting service classification. Simulation experiments have proofed the superior performance of PSOFEF. This Master thesis finally builds the routing model of service classification system using the experimental conditions of the project group and records the various business types of tha data drawing the business type map.The experimental result has proofed the superior performance of the routing model.