Design and performance evaluation of LTP enhancements for lossy space channels

Periodicals(2018)

引用 12|浏览6
暂无评分
摘要
AbstractSummaryThis paper deals with the performance of Licklider Transmission Protocol (LTP) in space environment, focusing on the impact of segment losses and the design of possible improvements. The in‐depth analysis of retransmission mechanisms carried out in the paper shows that, while LTP already achieves delivery time performance very close to the theoretical optimum in ideal conditions (no losses) and when losses affect data segments, a substantial improvement is still possible concerning signaling segments, whose loss can result in a significant increase of the LTP block delivery time. To this end, 2 simple enhancements, the Closing State and the proactive retransmission of signaling segments, are proposed and evaluated in the paper. Both enhancements have been added as optional features to the original Interplanetary Overlay Network LTP implementation (ION 3.5.0), in order to evaluate their effectiveness on a GNU/Linux testbed running the full protocol stack. Results presented in the paper confirm the validity of the proposed enhancements, as they significantly reduce the average delivery time in the presence of high losses and provide the additional advantage, even when losses are relatively low, of largely reducing the difference between average and worst cases.While LTP already achieves delivery time performance very close to the theoretical optimum in ideal conditions, a significant margin of improvement is still possible in the presence of high segment losses. Tests show that the performance improvement that the enhancements presented in the paper becomes rapidly huge with increasing loss rate. The enhancements have been introduced by NASA in the current ION version (3.6.x).Average delivery time vs SLR for various LTP enhancements. The enhancements significantly improve the average for SLR >5%. View Figure
更多
查看译文
关键词
delay-/disruption-tolerant networking, interplanetary networking, Licklider Transmission Protocol, performance evaluation, space communications
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要