On Price Differentiation In Internet Services Pricing And Resource Management

The advances of the Internet and information technology has enabled us to enter the "Information Economy" era.As the scale of Internet continues to grow,the demand of Internet services for network bandwidth and transmission speeds continues to increase.In addition,Internet applications and user preferences have diversified over the years.In light of this,Internet service providers are facing challenges such as bandwidth resource allocation,load balancing,and profits improvement.From a broader perspective,many large-scale systems also share network characteristics such as online electronics markets,data centers,smart grids,etc.These large-scale systems also face major challenges such as load balancing,profit optimization,resource allocation efficiency optimization,and matching demand with supply.Given the scale of these large-scale systems,traditional approaches based on protocols and rules exhibit high complexity and lack flexibility when in the face of changes.Therefore,the market has gradually become an important paradigm for large-scale system management and optimization.The market mechanism regulates the behavior of users in large-scale systems through price approaches to achieve load balancing and improve the efficiency of resource allocation/acquisition.Compared with traditional methods,market-based mechanisms are low in cost and flexible to adapt to changes,while at the same time aligning the interests of users with the global goals to a certain degree.Among market-based large-scale system management methods,price differentiation is a highly effective approach.By fully exploiting the differences in user characteristics,price differentiation set the prices of services or resources are set for different target user groups,so as to achieve load balancing,profit optimization,efficient resource allocation/acquisition,etc.This thesis studies the application of price differentiation in Internet services pricing and divisible resource management.In particular,we study two aspects:the optimal Internet service differentiation that maximize the profit of an ISP subject to bandwidth constraints;and the efficient allocation and acquisition of divisible resources(such as computing resources,power resources,etc.)in markets with network characteristics.For the former,in Chapter 2 we establish an optimal control framework to derive the analytical solution to an ISP's optimal service differentiation,i.e.,the optimal service qualities and associated prices.For the latter,in Chapter 4 we consider a market of divisible resource consisting of agents on both sides of demand and supply,and design an adaptive auction framework for a market mediator to achieve efficient resource allocation and acquisition by tuning price differentiation parameters.We also investigate the design of budget-balanced auctions for resource management in Chapter 4.We believe that our research on the optimal service differentiation of Internet service providers can help them design effective pricing strategies to increase profits,and guide regulators to regulate such practices so as to promote social welfare.Our work on resource management mechanisms can help large systems maximize the efficiency of resource allocation and supply through price diffentiation.Our theoretical models and research methods can be applied not only to the specific research scenarios in the thesis,but also have implications on the use of market-based approaches for congestion control,profit optimization,and efficient resource management in other large-scale systems with network characteristics.