The Study Of The Comunication Network Resource Based On The Pricing

Nowadays the broad band is increasing rapidly, but the Internet backbone network is not enough.The fibre-optical is enough, but the node equipment is lack.Inceasing the node equipmet requires amount of funds.The number of networking host is increasing, but the net profit model is not enough which leads to the result that the input of the operator is low. And now because of the limimted function of the network transport equipment, the network emerges the problem that is the network block which causes the transport efficiency is low. The number of networking host doubles every year, but the resource is not so abundant.At the same time, the network resource use presents the character which is application diversification and the commercialization.Especially in the village and the countryside, the person who uses the network resource is getting more and more. This dissertation studies the communication network resource distribution problem based on the pricing.By the way of pricing, the profit of the telecom enterprise can be increased and the users'utilization can be maximized.The research way of this dissertation includes the game theory, the system simulation, and so on.The dissertation builds the model of dynamic pricing and works out the optimal pricing in the case that the customers'demand situation is known.The model has important significance for many African countries where the Telecommunications industry is growing.The network telephone traffic and the cost are different in the different time and in the different CELL, and different customers have different elastic to the tariff price. The dynamic pricing may impulse and conduct the telephone traffic Capacity, increase the income of the Telecom Company, and improve the satisfaction degree of the customers.The demand function is founded.Then we can derive the revenue function. Then we can calculate the function's Maximum through matlab or mathmatics.This can make the profit of the telecom enterprise maximized.And then we can decline or increase the call price in different cell, so that the network call amount can be balanced. And then we give out the system implements of the dynamic pricing system, including the rate regulator, the block detecting regulator, and so on.A kind of dynamic pricing scheme is put forward.We present an integrated pricing architecture consisted of access price and usage price, which lead to an effective and real-time pricing method. With the method, different charging rates are applied to different service levels.The access price reflects the price of the flow reserved but not used the usage price which reflects the service. Difference introduces dynamic characteristics and changes cyclically as the network congestion state. This scheme urges users to select suitable service level and alternate users'ending rate according to the price changes through intermediate agent, thereby maximize utilities; besides those, ISP can interfere to help distribute customers among different levels.Such an integrated pricing mechanism in which supply and demand dominates,and plan assists guarantees the network to be utilized effectively. With the diffserv's feature, present a specific implementation framework. The framework designs on the diffserv model,combined with the dynamic pricing scheme we proposed, resolved the realization problems in the past.It monitors network congestion status through information sharing between the edge routers and consider this information to update prices.Congestion monitoring as well as the calculation of prices bases on end-to-end, which reduce the network load and control expenses.Between ISP and users increases intermediate agent,which get information from users at longer time-scales, displaying the current price category to users,but updating prices with ISP at shorter time-scales resolved time granularity problem.Admission control and pricing for connection-oriented elastic traffic are concerned.TCP admission control is depicted as a multi-objective optimization problem with blocked ratio minimization and active connection duration maximization as objectives.Assuming the connection request arrivals are determinate and stochastic respectively, we discuss how to price TCP connections.Then a dynamic pricing algorithm is examined which approaches the maximum of active connection duration while assuring the blocked ratio at acceptable level.Furthermore a price negotiation scheme based on Socks protocol is presented which can bring the pricing algorithm into practical networks.The equilibrium biding strategy and reserve price are discussed.Every user's strategy-proof strategy is to announce the true in the mechanism. The higher priorities are allocated to delay-sensitive users, and as a result, the total delay cost of all users is minimized.Rate control and pricing for connectionless elastic traffic become the topics of our interest.A conjugate gradient-like pricing algorithm is presented to adjust the network link prices in distributed manner, which shows faster convergence than gradient project algorithm.Then the probabilistic end-to-end price conveying and estimation method is analyzed formally, in which the choice of sampling size is of importance.An adaptive algorithm for sampling size adjustment is presented to assure the precision and creditability off price estimation.After that, the ideology of multi-scale congestion pricing is discussed which combines the admission control at connection level and rate control at packet level together.A production-consumption model for network resource allocation is presented.