SoylentNews is people
SoylentNews
Elon Musk says SpaceX Starlink satellite broadband beta testing starts in a few months:
This week [SpaceX] launched another batch of 60 satellites to bring the total size of its growing Starlink broadband constellation to more than 400. While it has the go-ahead to launch more than 12,000 satellites in the coming years, Musk said Wednesday that a "private beta" test of the service will begin in about three months, followed by a public beta about three months later for testers at northern latitudes.
In response to a Twitter user, Musk said Germany qualifies as far enough north, which could mean that much of northern Europe, Canada and the northernmost parts of the US may be eligible to try the service.
There is only so much bandwidth per satellite, so your pizza-box-sized transceiver would experience more congestion and lower throughput in an urban area than it would in a rural setting.
How many Soylentils are interested in signing up?
(Score: 2) by legont on Friday April 24 2020, @10:18PM (9 children)
Latency
"Wealth is the relentless enemy of understanding" - John Kenneth Galbraith.
(Score: 4, Informative) by takyon on Friday April 24 2020, @10:21PM (8 children)
~20-40 ms
https://en.wikipedia.org/wiki/Starlink#Satellite_hardware [wikipedia.org]
https://www.zdnet.com/article/elon-musk-spacexs-internet-from-space-should-be-good-enough-for-online-gaming/ [zdnet.com]
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 0) by Anonymous Coward on Friday April 24 2020, @10:51PM (7 children)
You expect the total trip latency to be less than or or equal to the earth-to-sat latency? (20-40 vs 25-35) I'm pretty sure you need to account for the return leg to reach the server/service you're measuring latency too.
Isn't it like saying my home internet is awesome since I can ping my router in 1-2ms? Or for the non-car-analogy-impaired: my gas pedal goes from 0-floored in 1 second?
(Score: 2) by takyon on Friday April 24 2020, @11:24PM (1 child)
A lot of numbers have been thrown around, including lower than 20. We'll just have to see what the beta testers say about it.
The main point is that it's nothing like a geostationary sat. It will be comparable to a normal ISP. On one of the previous stories, someone still thought that the latency was going to be around 1000 ms.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 2) by Runaway1956 on Saturday April 25 2020, @02:55AM
The latency (when it is finally measured) needs to be compared to today's latency in Outback, Nowhere. On some days, I can ping Google and get numbers in the 50's. Other days, I can't get anything under 100. It just sucks sometimes.
(Score: 4, Informative) by Immerman on Saturday April 25 2020, @12:09AM (4 children)
Your earth-sat latency is WAY off, characteristic of a pretty arbitrary orbit. A one-way latency of 30ms equates to about 9,000km altitude: about a quarter of the way to geostationary, and 16x higher than Starlink.
Starlink orbits at 550km, meaning the worst-case vertical round-trip signal latency is roughly 2*550km/(300e6m/s) = 3.7ms (A significantly diagonal transmission would nullify much of that thanks to the Pythagorean Theorem)
Even at their original planned altitude of 1000-1300km you're still under 8.7ms of vertical lag.
Meanwhile, their orbital signal path can actually do a fairly decent approximation of a minimum-distance great-circle path, so should be dramatically shorter than most fiber/wire paths available to terrestrial signals between arbitrary endpoints.
As an added bonus, the speed of light in vacuum is roughly twice the speed of light in fiber, roughly halving the latency for the same path length.
Of course, that's only relevant to the raw transmission times - router lag within the satellites and base stations is an independent variable that we don't know yet.
(Score: 0) by Anonymous Coward on Saturday April 25 2020, @01:21AM (3 children)
Even at their original planned altitude of 1000-1300km you're still under 8.7ms of vertical lag.
I think it will depend a lot on where you are going.
For example if I am in the US and want to goto Europe on the the thing the latency may be decently higher. you 9ms up and 9ms down + 1-2ms on station. Plus at least 1-2 ms on both ends on station. We are looking at at least 20-25ms just in network unless you add in more hops (which is possible). Plus the round trip packets back (another 20-25ms), so 40-50ms ping times is not unreasonable. Go to an external network you are looking at at least 3-5ms more transfer rate depending on what they can put into their backhaul and who they peer with. Plus whatever the peer network has.
For 'normal usage' of the internet I would put my money on somewhere between 80-120ms at least. For what some people have now this is amazingly better. For some people it would be a step down.
For me I do not play 'twitchy' games. So extra latency will not affect me much. So I probably would not even notice it.
I am looking forward to the real numbers. It really is a game changer in the internet. Those little podunk towns with providers milking their customers will have some real competition. Competition almost always brings out better prices and maybe more innovation.
(Score: 2) by Immerman on Saturday April 25 2020, @05:53AM (2 children)
8.7ms *round trip*, as in up and down combined - it's only 4.4ms to get from the ground to a satellite overhead, or back down again.
But that's the original orbit - the newer 550km orbit is much closer, with only 1.83ms one-way lag, or 3.7ms up-and-down. Now if you're talking ping times that will double since there's an up-and-down trip on the way to the server, and another up-and-down trip on the way back, but I wanted to make sure we're clear there - to get data from A to B in one direction, going to a 550 km orbit adds, at most, 3.7ms of signal lag.
For your Europe-to-US example - the shortest distance (as the bird flies) is about 8000km. That's 27ms of light-speed lag in vacuum, versus over 54ms of light-speed lag in fiber. The extra 3.7ms of lag to get to orbit and back again reduces your overall lag by almost 23ms, just because light speed is so much faster in orbit.
(Score: 3, Informative) by anotherblackhat on Saturday April 25 2020, @02:53PM (1 child)
Quibble: The speed of light in a standard glass fiber is only 31% slower than in a vacuum. So for 8000km, it's 27ms vs. 39ms not 54ms.
(Score: 2) by Immerman on Saturday April 25 2020, @03:04PM
So it is. I must have misremembered.