跳转至

Delay is Not an Option: Low Latency Routing in Space

ABSTRACT

SpaceX has filed plans with the US Federal Communications Committee (FCC) to build a constellation(星座/星宿) of 4,425 low Earth orbit communication satellites. It will use phased array antennas for up and downlinks and laser communication between satellites to provide global low-latency high bandwidth coverage. To understand the latency propertes of such a network, we built a simulator based on public details from the FCC filings. We evaluate how to use the laser links to provide a network, and look at the problem of routing on this network. We provide a preliminary evaluation of how well such a network can provide low-latency communications, and examine its multipath properties. We conclude that a network built in this manner can provide lower latency communications than any possible terrestrial optical fiber network for communications over distances greater than about 3000 km.

SpaceX已向美国联邦通信委员会(FCC)提交计划,拟建造一个由4,425颗低地轨道通信卫星组成的星座。该星座将使用 相控阵天线进行上下行链路通信 ,并通过 激光通信实现卫星间的连接 ,以提供全球低延迟、高带宽的覆盖。为了理解该网络的延迟特性,我们基于FCC公开的文件构建了一个模拟器,评估了如何利用激光链路构建该网络,并研究了网络中的路由问题。我们提供了该网络能够提供低延迟通信的初步评估,并考察了其多路径特性。我们得出结论,使用这种方式构建的网络,在超过3000公里的通信距离上,能够提供比任何可能的陆地光纤网络更低的延迟

INTRODUCTION

alt text

As network bandwidths have increased, latency has emerged as being the limiting factor for many networked systems, ranging from the extremes of high frequency trading, to the more mundane effects of latency on VoIP, online gaming, and web performance[2]. Fundamentally, once traffic engineering has mitigated congestion[7, 9] and buffer bloat has been addressed, for wide-area traffic the remaining problem is that the speed of light in glass simply isn’t fast enough.

随着网络带宽的增加,延迟已成为许多网络系统的限制因素,从高频交易的极端情况,到延迟对语音通信(VoIP)、在线游戏和网页性能的日常影响[2]。从根本上讲,一旦流量工程解决了拥塞问题[7, 9],并且缓冲区膨胀得到了处理,那么广域网流量而言,其面临的剩余问题是:光在玻璃中的传播速度实在不够快。

In recent FCC filings[12], SpaceX proposed and subsequently received permission to launch Starlink, a constellation of low Earth orbit (LEO) satellites to provide low-latency, high-bitrate global Internet connectivity. These filings provide a great deal of detail about the RF links between the satellites and the ground, including how phased-array antennas can steer narrow transmission beams for both up and downlinks. The filings (备案) do not discuss in any detail satellite to satellite communications, but do state that free-space lasers will be used. No radio spectrum for satellite-to-satellite communication is requested, so lasers must be the primary communication link between satellites. Crucial to the low-latency story is that free-space lasers communicate at c, the speed of light in a vacuum, which is ≈ 47% higher that in glass[4].

在最近的FCC文件中[12],SpaceX提出并获得批准发射Starlink,这是一个低地轨道(LEO)卫星星座,旨在提供低延迟、高比特率的全球互联网连接。这些文件详细介绍了卫星与地面之间的射频链路,包括如何通过相控阵天线为上下行链路引导窄波束。文件中没有详细讨论卫星间通信,但指出将使用自由空间激光进行通信。由于没有申请卫星到卫星通信的无线电频谱,因此激光必须成为卫星间的主要通信方式。对于低延迟而言,至关重要的一点是自由空间激光的通信速度为光速 c,即真空中的光速,这比玻璃中的光速快大约47%[4]。

Starlink represents a new category of wide-area backbone, where thousands of satellites move and connect in a predictable pattern, but due to orbital constraints the network is far from a simple static mesh. We ground our study in the basic properties of the Starlink deployment, and proceed by simulating routing designs on such a network. Where details are not publicly available, we adopt reasonable parameters from first principles. Our goal is to provide early insight into the interactions between the dynamic topology of the constellation, how routing might work over such a novel network architecture, and emergent end-to-end latency properties.

Starlink代表了一种新型的广域骨干网络,在这个网络中,成千上万颗卫星以可预测的模式移动和连接,但由于轨道约束,网络远非简单的静态网格。我们的研究基于Starlink部署的基本特性,接着对这种网络架构上的路由设计进行模拟。在公开细节不足的情况下,我们从基本原理出发,采用合理的参数。我们的目标是为该星座的动态拓扑、路由如何在这种新颖网络架构上运行,以及由此产生的端到端延迟特性提供早期的见解。

What is Buffer Bloat

Bufferbloat 是指在网络中由于过度的缓冲导致的数据包延迟增加和网络性能下降的现象。当网络设备(如路由器或交换机)配置了过大的缓冲区时,数据包在这些缓冲区中长时间排队,从而引发高延迟和抖动,特别是在需要低延迟的应用场景中,如VoIP、在线游戏和视频通话等[1][4][8]。

Bufferbloat的工作原理

在一个理想的网络中,数据包应该被迅速传输到目的地。然而,当网络流量突然增加时,设备的缓冲区可能会被填满,导致新的数据包无法立即处理。这些数据包会在缓冲区中排队,等待前面的数据包被处理完成。由于TCP协议的特性,发送方会继续发送数据,直到缓冲区满为止,这样就形成了一个正反馈循环,使得延迟不断增加[2][5][11]。

例子解释

想象一下一个繁忙的餐厅,顾客(数据包)进来后需要在前台(路由器)登记。假设前台有一个大桌子(缓冲区),可以容纳很多顾客。虽然这个大桌子可以容纳许多顾客,但如果餐厅的服务员(后端处理能力)只能每次处理一位顾客,那么即使有很多顾客在等,也无法快速为他们提供服务。

  • 正常情况:顾客进来登记后,可以很快被服务员接待,整个过程流畅。
  • Bufferbloat情况:当大量顾客同时进来登记时,他们会在前台长时间等候。即使前台有足够的空间容纳这些顾客,由于服务员处理速度慢,顾客仍然需要等待很长时间才能被接待。这种情况下,即使餐厅的座位很多,但由于服务员忙不过来,整个用餐体验却变得很糟糕。

在网络中,这种情况表现为高延迟和不稳定的连接,即使带宽很高,也无法保证良好的用户体验[3][6][19]。

总结

Bufferbloat是一个严重影响网络性能的问题,尤其是在需要低延迟的应用中。理解这个现象有助于设计更有效的网络架构和优化TCP协议,以减少延迟并提高用户体验。解决Bufferbloat的方法通常包括优化缓冲区大小、使用主动队列管理(AQM)算法等,以确保数据包能够及时处理,而不是在缓冲区中长时间等待[4][22][24]。