INTRODUCTION

Low-earth-orbit (LEO) satellite networks (LSNs) are evolving rapidly in recent years, thanks to the fast deployment of mega-constellations such as SpaceX’s Starlink [10], Eutelsat OneWeb [5] and Amazon Kuiper Project [2]. Emerging LSNs aim to provide broadband coverage and low-latency services globally, and are carrying an increasing amount of Internet traffic [3]. For example, Starlink, the largest operational LSN today, has attracted more than 4 million worldwide customers on 7 continents as of September 2024 [7].

低地轨道（LEO）卫星网络（LSNs）近年来发展迅速，得益于诸如SpaceX的Starlink 、Eutelsat OneWeb 和亚马逊Kuiper项目等巨型星座的快速部署。新兴的LSNs旨在全球范围内提供宽带覆盖和低延迟服务，并承载着日益增长的互联网流量。例如，截至2024年9月，Starlink作为目前最大的运营LSN，已吸引了遍布七大洲的超过400万用户。

One key property differentiating LSNs with other existing terrestrial networks is that: a portion of the network infrastructure (i.e., the space segment) is moving at a high velocity related to the earth surface. Hence, for a network path through LSN, a sub-set of its intermediate links continuously change over time, imposing substantial challenges on existing end-to-end congestion control algorithms (CCAs).

与其他现有地面网络相比，LSNs的一个关键特性是：部分网络基础设施（即空间段）相对于地球表面以高速移动。因此，对于通过LSN的网络路径，其部分中间链路会随着时间不断变化，这对现有端到端拥塞控制算法（CCAs）提出了重大挑战。

On the one hand, the unique LEO mobility not only results in rapidly varying satellite link capacity, but also involves frequent non-congestion RTT variations (e.g., caused by path fluctuations) and bursty packet losses (e.g., due to groundsatellite handovers). On the other hand, existing end-to-end CCAs leverage the network performance changes observed on the sender to infer congestion, but they are unable to discriminate whether the network variation is exactly caused by a congestion event (e.g., queuing at the bottleneck link) or not. As we will show in §3, the frequent non-congestion network variations caused by LEO mobility often mislead existing CCAs such as TCP Cubic [23], Vegas [13], Copa [11] etc., and further result in self-restrained CCA performance.

一方面，LEO独特的移动性不仅导致卫星链路容量快速变化，还涉及频繁的非拥塞RTT变化（例如，由路径波动引起）以及突发性数据包丢失（例如，由于地面与卫星之间的切换）。另一方面，现有的端到端CCAs通过发送端观察到的网络性能变化来推断拥塞，但无法区分网络变化是否确实由拥塞事件（例如瓶颈链路排队）引起。如我们将在§3中展示的，由LEO移动性引起的频繁非拥塞网络变化通常会误导现有CCAs，例如TCP Cubic 、Vegas 、Copa 等，并进一步导致CCAs性能受限。

Note

LEO的高速移动性 -> 性能指标改变，但是原因不是拥塞，而是高速移动产生的副作用（“非拥塞网络变化”） -> LEO CC 认为拥塞发生并进行“本毫无必要”的调整 -> 链路利用率低 / 延迟高

We advocate that CCAs in LSNs require more effective indicators that can help the sender discriminate non-congestion performance changes and estimate network conditions more accurately. On our further investigation (§4), we find that the unique LEO mobility is managed and scheduled by a special feature called "LEO reconfiguration" in LSNs, which is closely related to the non-congestion network variations, and thus is a potential indicator for discriminating non-congestion network variations (e.g., link capacity and propagation RTT changes due to LEO mobility).

我们主张，LSNs中的CCAs需要更有效的指标，以帮助发送端区分非拥塞性能变化并更准确地估计网络状况。在进一步研究中（§4），我们发现LEO独特的移动性通过一种特殊功能——“LEO重配置”在LSNs中进行管理和调度。该功能与非拥塞网络变化密切相关，因此可能成为区分非拥塞网络变化（如链路容量和由于LEO移动性导致的RTT传播变化）的潜在指标。

Note

我们需要一个“衡量因子”作为指标，帮助 发送端 区分 “这”是不是“非拥塞网络变化”

好在我们找到了一个因子，叫 LEO reconfiguration，它跟LEO运行轨迹等有关

To demonstrate the potential benefit of LEO reconfiguration, we conduct case studies and modify the congestion control logic of two representative CCAs: Copa and BBR (§5). Our preliminary results based on trace-driven network emulation show that even with a simple reconfiguration-aware modification: (i) Copa achieves 74.4% throughput improvement; and (ii) BBR reduces 30.2% delay on average, as compared to their original implementations. We finally discuss several open questions and hope they can inspire future CCA research for emerging LSNs (§6).

为了展示LEO重配置的潜在优势，我们进行了案例研究，并修改了两种具有代表性的CCA（Copa和BBR）的拥塞控制逻辑（§5）。基于轨迹驱动的网络仿真，我们的初步结果表明，即使仅进行简单的重配置感知修改：（i）Copa实现了74.4%的吞吐量提升；（ii）BBR平均减少了30.2%的延迟，相较于其原始实现。

最后，我们讨论了一些开放性问题，希望能为新兴LSNs中的CCA研究提供启发（§6）。