| Real-time multimedia applications require either guaranteed or targeted Quality of Service (QoS). This can be achieved by dynamic resource allocation mechanisms in reservation-based QoS architectures or by traffic dropping and scheduling mechanisms in priority-based QoS architectures. Here, we investigate the feasibility of dynamic network resource and QoS control by dynamically adjusting prices based on congestion.; We first consider a pricing policy in a reservation-based QoS architecture, which implements a distributed resource allocation to provide guaranteed bounds on packet loss and end-to-end delay for real-time applications. Distributed pricing roles are assigned to each user, each network node, and an arbitrager in between the user and the network. We show that a resource allocation is optimal if all users, arbitragers, and network nodes are in equilibrium. When delay constraints are not binding, we investigate two dynamic pricing algorithms using gradient projection and Newtons method to update prices, and prove their convergence. We analyze the performance of the dynamic pricing policies and show that the gradient algorithm using Newtons method converges more quickly and displays only a few small fluctuations. When delay constraints are binding, we investigate subgradient methods which can provide convergence to some range of the optimal allocation.; We then turn to study a pricing policy in a priority-based QoS architecture and investigate whether congestion-based pricing can be used to control aggregate traffic into each codepoint by motivating users to choose the codepoints appropriate for each application. Distributed pricing roles are assigned to each user and each autonomous system. We show that a traffic allocation is optimal if and only if all users and autonomous systems are in equilibrium, and that minimal information exchange between users and autonomous systems including QoS measurements and marginal utility is required. We then investigate the relative precisions of QoS, QoS sensitivity, and utility sensitivity. We found that the corresponding relative precisions of prices, rates, and utility are all inversely proportional to the square root of the total number of observations of QoS measures. As a result the pricing policy proposed is feasible only for networks with sufficiently high bandwidth to guarantee that QoS can be quickly measured. |
