Study Of Pricing Strategy For Differentiated Services With QoS Constraints In Communication Networks

With the rapid development of the network and the increasing demand of the network resources, the network traffic is growing explosively. As a result, the contradiction between the relatively limited network resources and the growing customer demand is serious. Technical means alone have been unable to solve this problem, so people are turning to the economic approach., and many pricing strategies have been proposed. Traditional study on pricing was usually based on the IP network infrastructure, which means the network provides best-effort service to costomers. However, with the continuous development of network services, some streaming media applications such as Voice over IP, video-on-demand emerge. These applications often have a certain quality requirements on bandwidth, loss rate or delay. We aim to find a pricing strategy that can differentiate services according to their types and meet different services’ QoS requirements.We aim to propose a differentiated service pricing strategy that can achieve three goals. First, QoS guarantees. Namely to provide services accords to the QoS requirements and fulfills services’requirements on transmission rate, missing data probability, delay and delay jitter. The second is to achieve service differentiation. ISP should differentiate services according to heterogeneous requirements of the customers and price differently. Finally, we aim to construct a pricing model to achieve the social welfare maximization.To achieve the goals, we need to solve three problems:QoS constraints, to service differentiation and utility. There are already a variety of solutions to provide QoS guarantees such as Integrated Service (IntServ) mechanism and Differentiated Services (DiffServ) mechanism. IntServ provides guaranteed service from the perspective of resource reservation but it has bad scalability and high complexity. Therefore study on IntServ-based service pricing is less. DiffServ ensures the ability to differentiate services, but only a limited number of service levels is provided, which cannot meet the arbitrary requirements of service. In addition, we used to describe QoS from the engineering point and the description used for pricing is less studied. We expect to find an accurate QoS description suitable for pricing.Effective bandwidth provides a new way to quantify QoS and differentiate services. In asynchronous network, multiplexing is used to improve resource utilization. As a result, the concept "Effective Bandwidth" is proposed to measure the usage of network resources in the network that allows statistical multiplexing. We can quantify QoS using effective bandwidth. At the same time, due to the presence of the background flow, the network capacity cannot be fully allocated to users, which influences QoS of services. To provide stable and accurate QoS, we need to quantify the influence of background flow in QoS description. Pricing based on network utility model needs to be built on the assumption that we know the utility function, so it has poor applicability. Fortunately, we find that elastic UE issues can express the social welfare maximization problem, which avoids the assumption of the utility function.We review the background and what is studied recently of service differentiation and QoS pricing, analyze the existing problem of these study in this paper. We emphasize on QoS and its parameters, and give the QoS definition through effective bandwidth. The relation between QoS and effective bandwidth is given and the impact of background flow is quantified to the formula. Pricing model of service differentiation based on QoS constraints is proposed. It can achieve service differentiation and QoS guarantee. Besides, Multiplexing and background flows are considered in this model. Finally, by analyzing the elastic UE, we propose to solve the communication pricing problem through the elastic UE problem, which solves the uncertainty of the utility function.