Starlink revisions, Nov 2018

> This might as be considered a law of physics - electromagnetic energy will always travel faster, more... through a conductive medium than through air. it actually travels faster through air (no vacuum necessary) than through fiber or copper as those transmit information at 2/3rd of the speed of light while air transmits it at 99% [1]. which is why starlink could in theory achieve better latency than a straight run of fiber optic across the world [2]. this of course doesn't mean that it's better in a mouse because of the overhead of processing such a short distance. [1] link [2] ▶ this video

Mark Handley has done a bunch of videos analysing this on his channel[0]. The TLDR is that there is a lot of room for Starlink to beat existing terrestrial options IF they can get the laser links working on-orbit. This video[1] shows estimates using what was known about proposed configurations in 2018 [0] link [1] ▶ this video

Forward and back looking lasers would have minimal speed difference, side laser links don't appear to be at extreme speed differences according to these hypothesized connections: ▶ this video

▶ this video is an animation of how it's supposed to work, including latencies There are enough satelites and low enough routing on the lasers to mean packets will arrive from London to NY faster than even a great-circle fibre. In reality how much bandwidth is available is a function of money, and HFTs tend to have a lot of money.

There's no doubt that this sort of technology can in principle deliver sub-100ms latencies. Take a look at my analysis here with inter-satellite links: ▶ this video And here without ISLs: link Now, these videos assume no queuing delay. It's really hard to guarantee low queuing delay in a traditional IP network, but there's been a lot of research over the years, and we have a range of ideas that can be used to deliver minimal queuing with reasonable utilization. In fact doing this for Starlink is something I'm actively researching at the moment. I'm convinced it can be done, but it won't look like a conventional IP network internally. I don't know what SpaceX will actually do, so if the FCC has doubts, this is perhaps where they are concerned. Other delays might be in the ground segment, but that's more or less the same for SpaceX's terestrial competitors, and the FCC seems to think they can do it.

If you want a really great visualization and latency estimates for starlink check out the models and videos this guy came up with: ▶ this video link The comment from the FCC about it not accounting for "processing [&] routing" is kindof nonsense on its face, those are not novel problems they're already done by ever other router and switch, including several other satellite networks.

At 5:03 [1] he talks about how the laser links give good connectivity east-west but not as good north-south. [1] ▶ this video

Here's some great visualizations of the orbits and an explanation of what it means that the orbits are optimized for the northern hemisphere. (This confused me at first since obviously any orbit has to go just as far south as it does north of the equator.) ▶ this video

this guy made an excellent youtube video modeling and visualizing the satellite orbits, the network routing paths across them, and the resulting latencies: ▶ this video NYC to London on the current internet is 76ms, he shows it across starlink's updated plans as working out to around 46ms. Later in the description he points out a mistake in his own math and says that starlink paths should be 8.7% faster than the video shows, so really jfk<->lhr ~ 42ms.

The FCC filing has a lot of good technical details on the deployment and frequency usage. Additionally, Mark Handley has a good primer on it. It may kill actual east-west terrestrial traffic due to latency improvements: ▶ this video North-South will still be terrestrial.

There were recently two simulations done on possible latencies based on the FCC filings. You can find the video simulations in the links at the bottom of this post. You can see international links that have latencies lower than the existing internet by a significant margin. Unfortunately neither of those simulations take into account that the first iteration has been changed to only use radio and no laser links. [0]: link [1]: ▶ this video