Satellite networks are booming to provide high-speed and low latency Internet access, but the transport layer becomes one of the main obstacles. Legacy end-to-end TCP is designed for terrestrial networks, not suitable for error-prone, propagation delay varying, and intermittent satellite links. It is necessary to make a clean-slate design for the satellite transport layer. This paper introduces a novel Information-centric Hop-by-Hop transport layer design, INTCP. It carries out hop-by-hop packets retransmission and hop-by-hop congestion control with the help of cache and request-response model. Hop-by-hop retransmission recovers lost packets on hop, reduces retransmission delay. INTCP controls traffic and congestion also by hop. Each hop tries its best to maximize its bandwidth utilization and improves end-to-end throughput. The capability of caching enables asynchronous multicast in transport layer. This would save precious spectrum resources in the satellite network. The performance of INTCP is evaluated with the simulated Starlink constellation. Long-distance communication with more than 1000km is carried out. The results demonstrate that, for the unicast scenario INTCP could reduce 42% one-way delay, 53% delay jitters, and improve 60% throughput compared with the legacy TCP. In multicast scenario, INTCP could achieve more than 6X throughput.
翻译:卫星网络正在蓬勃发展,以提供高速低纬度互联网接入,但运输层已成为主要障碍之一。传统端到端TCP是为地面网络设计的,不适合容易出错、传播延迟和间歇性卫星连接。有必要为卫星运输层设计干净的版面设计。本文介绍了一个新的信息-以Hop-Hop-Hop-Hop-Hop-Hop 运输层设计,INTCP。它利用缓冲和请求响应模型,执行跳式包的再传输和跳式拥堵控制。希望式的端到端TCP为地面网络设计,不适合容易出错、传播延迟和间歇性卫星连接。需要为卫星运输层设计一种以信息为中心、以Hop-Hop为中心、Hop-Hop-Hop-Hop-Hop为主的运输层设计。它可以节省卫星网络的珍贵频谱资源。在模拟星联星座星座中评估INTCP的性能性能。通过1000公里以上的远程通信回收袋回收,减少再传送延延延延延延延。INCT的频率为4000%,比延延后推延后推延延后,将延后推延后推延后推延后推延后推延后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后推后