[Starlink] speedtest.net takes a look at sat internet around the globe

[Darrell Budic]

This matches what I’ve seen from my side. I host a speedtest.net server at the $dayjob in Chicago, with ipv6. That network is peered with Google's, so all of it’s 10G connection is available to it. Of course, it’s way down the list in speedtest servers for me because of the geographic nature of speedtest.net’s recommendations. Those recommended servers are even more useless with a service like starlink, but that doesn’t seem to bother speedtest. 

With v6 over starlink, it does make it a reasonable test of throughput for me, since I have a better than usual understanding of the path and available bandwidth in my test case. Also why I suspect I’m most of the data for Iowa County, WI, I’ve run a fair number of speedtests even without automation :)

  -Darrell

> On Aug 12, 2021, at 6:24 PM, Ricky Mok <cskpmok at caida.org> wrote:
> 
> I think the blog used the data they collected from speedtest.net. I conducted research on that topic. Ookla's speedtest is TCP based (usually 4-6 concurrent TCP flows to a test server). Ookla deployed over 12k test servers (many ISPs contributed to it). The test used IP geolocation to pick a set of ten nearest servers geographically and establish WebSocket to send "ping"s to test the network latency. The test then chooses the server with the lowest latency and send 10 more probes to test the RTTs. The "latency" they referred should be the extract from the RTT tests.
> 
> I think it is a norm to have shared last mile and advertise the "max" capacity, just like what cable providers did.
> 
> I remembered when they start launching the beta test. Only that latitude range had continuous satellite coverage. I assume the service outside that range would be unstable. But they have ground station built out of that range already. So, they just need to launch more SpaceX rockets...
> 
> For CDN, they signed agreements with Google (and Azure?) I expect ground stations have direct interconnection with the cloud providers for content. They can also have caches installed in the ground station (e.g., netflix, youtube), similar to some ordinary ISPs installed in-network caches in Central Offices. However, i agree that they cannot do caching on the satellites, particularly we have HTTPS :)
> 
> Ricky
> 
> On 8/12/2021 3:34 PM, Ulrich Speidel wrote:
>> It always pains me to see "speed" tests like these, especially if the methodology they've used isn't clear. My big gripes:
>> 
>> - I strongly suspect that the speed tests here (and in so many other blogs/vlogs) are UDP-based, which doesn't tell me a thing about how much TCP goodput I'll get over any of these links.
>> - Latency is measured between end user and ... what? The satellite gateway? Some imaginary fixed point on the Internet that all our traffic has to to through? Or maybe just speedtest.net's servers, whose locations possibly don't matter one iota for my Internet performance?
>> - If we reasonably assume that the capacity of a Starlink satellite needs to be shared between its users, then few users / satellite equates to a large share of the capacity. From Starlink's front page: "Starlink is available to a limited number of users per coverage area at this time." Guess what? What we see here may not last, but it's sure great for marketing.
>> - Ever wondered why Starlink's bulk of beta users sits between 40-something and 50-something degrees of latitude? That's right, because that's where you get the largest concentration of satellites right now, which helps keep the number of users per satellite down. Elsewhere? Tough luck.
>> - At the other end of your terrestrial broadband connection might be a few CDN servers, meaning you and your fellow customers will only need to use the ISP's feed once for that viral cat video everyone wants to watch. Starlink goes direct to site, not to a local ISP. So if your ISP is in space and the CDNs are on the ground, a thousand Starlink users on a satellite wanting to watch the cat video will need to bring it across the satellite a thousand times. Your terrestrial ISP (or even sat-based ISP with a terrestrial network connecting end users) only needs to do this once.
>> 
>> But big numbers always look great, don't they?
>> 
>> On 13/08/2021 9:22 am, Darrell Budic wrote:
>>> https://www.speedtest.net/insights/blog/starlink-hughesnet-viasat-performance-q2-2021/ <https://www.speedtest.net/insights/blog/starlink-hughesnet-viasat-performance-q2-2021/>
>>> 
>>> Nothing we didn’t know, but interesting comparisons between the 3 sat companies and fixed breadboard around the world.
>>> 
>>> Made me wonder if there’s anyone else contributing to the speed tests in Iowa county, WI, looks a lot like my averages there...
>>> 
>>> 
>>> _______________________________________________
>>> Starlink mailing list
>>> Starlink at lists.bufferbloat.net <mailto:Starlink at lists.bufferbloat.net>
>>> https://lists.bufferbloat.net/listinfo/starlink <https://lists.bufferbloat.net/listinfo/starlink>
>>> 
>> -- 
>> ****************************************************************
>> Dr. Ulrich Speidel
>> 
>> School of Computer Science
>> 
>> Room 303S.594 (City Campus)
>> Ph: (+64-9)-373-7599 ext. 85282
>> 
>> The University of Auckland
>> ulrich at cs.auckland.ac.nz <mailto:ulrich at cs.auckland.ac.nz> 
>> http://www.cs.auckland.ac.nz/~ulrich/ <http://www.cs.auckland.ac.nz/~ulrich/>
>> ****************************************************************
>> 
>> 
>> 
>> 
>> 
>> _______________________________________________
>> Starlink mailing list
>> Starlink at lists.bufferbloat.net <mailto:Starlink at lists.bufferbloat.net>
>> https://lists.bufferbloat.net/listinfo/starlink <https://lists.bufferbloat.net/listinfo/starlink>
> _______________________________________________
> Starlink mailing list
> Starlink at lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/starlink

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.bufferbloat.net/pipermail/starlink/attachments/20210815/cf2ad93e/attachment.html>