Haipeng Jin

Performance comparison of header compression schemes for RTP/UDP/IP packets

Multimedia over wireless is one of the major applications envisioned for the next generation all-IP wireless networks. To efficiently use the scarce radio resources, header compression schemes are proposed to remove redundant information in the RTP/UDP/IP headers and yet still provide reliable communications over the error prone wireless links. In this paper, we study and compare the performance of two of the most popular header compression schemes: enhanced compressed RTP (ECRTP) and robust header compression (ROHC). Our simulation results show that the packet error rate (PER) performance of ECRTP and ROHC is similar to each other when the local repair mechanism in ROHC is not considered. The only exception is ROHC U mode, where the PER is much worse. All ROHC modes have lower average header size and overhead than ECRTP. In particular, ROHC R mode offers the lowest average header size and overhead in the cases we studied.

Performance of a tree-based collision resolution algorithm for random access with smart antennas

A smart antennas ability to simultaneously resolve more than one user on the same channel is exploited to help expedite the process of random access, especially for reverse channel signalling. However, the use of a smart antenna requires frequent array adaptation, a difficult task in a random access environment. In our paper of 2003, we proposed and studied the SINR performance of a media access protocol intended for use with smart antennas. In this paper, we study the random access performance of the proposed protocol when a tree-based fast collision resolution algorithm is used. The maximum achievable throughput and an upper bound on the expected delay are found. The impact of the number of antennas on the performance with both flat fading and frequency selective fading is studied. Typical results show significant improvement in throughput for systems with smart antennas. We also show that dynamically adjusting the retransmission probability in the collision resolution process is not worthwhile since this extra complexity produces a gain of only about 6.2% in the best case.