This paper describes our experiences in the development of the UDP-based Data Transport (UDT) protocol, an application level transport protocol used in distributed data intensive applications. The new protocol is motivated by the emergence of wide area high-speed optical networks, in which TCP is often found to fail to utilize the abundant bandwidth. UDT demonstrates good efficiency and fairness (including RTT fairness and TCP friendliness) characteristics in high performance computing applications where a small number of bulk sources share the abundant bandwidth. It combines both rate and window control and uses bandwidth estimation to determine the control parameters automatically. This paper presents the rationale behind UDT: how UDT integrates these schemes to support high performance data transfer, why these schemes are used, and what the main issues are in the design and implementation of this high performance transport protocol.

We discuss scheduling policies in wireless networks with fading channels. A utility function that takes both the communication throughput and the transmission delay into account is defined. A dynamic programming optimization method is used to obtain the optimal scheduling policy which explores the channel dynamics and gets a reasonable trade-off between the communication throughput and the packet transmission delay. A near-optimal solution is also presented to deal with problems with large state space. Through simulation results, we show that our policy can improve the performance in both aspects of throughput and transmission delay, compared to the well-known weighted round-robin scheme which does not consider channel dynamics.

As one of the most demanding applications in high speed networks, real-time services, such as audio and video, have stringent requirements on quality of service (QoS) guarantee, especially the bounded end-to-end delay and delay variations. Accordingly, the real-time services with deterministically bounded delay have been widely studied. The key is to design efficient scheduling algorithms. In this paper, a scheduling method, referred to as static-rotating-priority-queues (SRPQ), is proposed. Exact schedulability conditions for the method, which is important for call admission control, is also presented. The most desirable features of this method are its low complexity and reduced number of first-in-first-out (FIFO) queues in a scheduler while providing high bandwidth utilization.

In this paper, an analytical model is developed for the study of the impact of user behavior on the effectiveness of pricing, service quality, and link bandwidth allocation/sharing in a multiple class-of-service Internet. The model characterizes a variety of user behaviors, such as user sensitivity to service price/quality and user migration from one class of service to another. It also incorporates a usage-based pricing scheme and a general trunk reservation policy. In particular, a three-tier service model for a single link is considered. Numerical studies reveal the richness of this research issue and provide insights into the optimum selection of service price, quality and trunk reservation values with given user behaviors. We find that trunk reservation is effective in terms of service protection only when the call migration effect is small and the call migration effect can be harnessed by a proper design of service pricing/quality structure. We also find that user sensitivities have great impact on service price/quality structure design and care must be taken to avoid sudden network performance degradation at certain price/quality parameter values.