A Survey Of IP QoS And Network Utility Maximization

To study the effectiveness of resource allocation based on utility has been widely used, this method use the Utility Function to study the issues of resource allocation. Utility Function refers to a function which describes the degree of user satisfaction with a certain amount of allocated resource to user. NUM (Network Utility Maximization) as a network model is widely used in central control network, it relate the users’satisfaction to allocation schema. The processing of the NUM framework firstly must build the network utility maximization mathematical model and then use optimization theoretical to solve the model. IP QoS (Quality of Service) is a control mechanism, which is used to measure quality of service. IP QoS service model relate to aspects of network service, such as error rates, bandwidth, throughput, transmission delay, availability, jitter and etc. IP QoS have been more and more important in this years as three type of network begin to amalgamation. How to combine network utility maximization with IP QoS, to provide users with efficient and diverse network services, while ensuring scalable network traffic type. Scalability requirements can be more flexible configuration user traffic types, efficient requires a network using the lowest and cost to provide satisfactory service to users.This paper study theoretical model of IP QoS and NUM, discussing the various of IP QoS model and differences of common queue management strategy, meanwhile, describes the theoretical model and optimization method of NUM model. Then we study the relationship of IP QoS and NUM, combine theoretical model of QoS with NUM. Then I propose using parameters to distinguish different types of traffic (such as Best Effort, Soft QoS, Hard QoS), give the explanation of different of parameters to different traffic. Finally, using unified utility function to build the NUM model, after that, I propose the reasonable assumptions based on the practice of life, after that, I solve the problem using optimization theory and design algorithm to allocate the resource to users. At the end of this, we use two group of experimental data to simulate the algorithm. This algorithm can achieve the target that the user can obtain both the defined quality of service and maximize the utility of the network.