Characteristics of Resource Allocation using Pricing

We consider pricing of network resources in a reservation-based QoS architecture. The pricing policy 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 explore the characteristics of the optimal resource allocation by comparing it with a simpler method based on traffic characteristics alone. We show that pricing based optimal resource allocation can provide higher network utilization by allocating more network resources to classes that are more sensitive to additional network resources. We give sufficient conditions to establish that an increase in the price for a resource results in a decreased demand for that resource, an increased demand for the other resource at that node, and an increased demand for resources at all other hops. Finally, we provide two dynamic pricing policies using simple proportional step and gradient feedback laws. We analyze the performance of these two approaches, and show that the gradient algorithm converges more quickly and displays only a few small fluctuations.