[pocketstar]

No spacecraft can hope match terrestiral latency, the speed of light is not your friend. There was an identical push for space ISPs in the 90's. Iridium was the only sucessful leo space "ISP". Elon has been sucessful with what he says he will do but all these space bases "ISPs" are vaporware until they deliver. Look at the joke of oneweb with their botched softbank merger with intelsat, consilidating billions of dollars of existing debt to fund another few billion to develop a new constillation does not bode well.

Echostar is doing a good job with their proven GEO ISP constillation. With several billion in the bank they are the most promising to be sucessful, keep an eye on their HTS development over the next couple years. The problem with GEO ISPs is they have 1-2s latency :(

[mikeash]

Latency is not a problem for low orbit. You're adding a few hundred miles to the trip, which adds single-digit milliseconds of latency. Cost is a huge problem (that's why Iridium went bankrupt) but SpaceX will be able to launch their constellation far more cheaply than it cost Iridium in the 90s.

I agree that they're vaporware until they deliver, but latency isn't going to be the killer.

[vonklaus]

SpaceX was the launch provider for a 10-sat Iridium cluster 3 weeks ago so they must have restructured favorably as that launch succeeded on May 2, 2017.

[mikeash]

The original Iridium was viable when it came to operating costs, they just had no hope of success when they also had to pay back the ~$6 billion in capital costs it took to build and launch the satellites.

The current incarnation of Iridium bought the constellation for only $35 million, effectively discharging the massive debt and allowing them to profitably operate the satellites without having to pay back what it cost to put them up in the first place.

Come forward to today and satellites are cheaper and more capable, demand for communications is higher, and launch services are much cheaper, so Iridium thinks they'll be able to put up a next generation constellation without repeating their history. It looks like they're planning on spending about $2.1 billion to build the satellites and $800 million to launch them, which puts the new constellation at about one third the cost of the original taking inflation into account, while being vastly more capable. (The current Iridium system gets you a connection which provides either a single voice line or 2.4kbps (!) data service. The next generation will go up to 1.5Mbps.)

Iridium is a fascinating story of hubris, spectacular failure, and eventual success.

Unrelated fun fact, an Iridium satellite was involved in the first accidental collision of two satellites. Iridium 33 collided with the defunct Kosmos-2251 in 2009. The relative speed was over 26,000MPH and sprayed a bunch of debris around the two orbits. Iridium has spares in orbit, so they were able to patch up the constellation without much trouble.

[vonklaus]

SpaceX carried an Iridium cluster 3 weeks ago. OneWeb, as you noted; also exists. Google has designs in the space and you can see a company like Planet (or even Planet) eyeing the economics of this. My point was that there are companies putting pressure on terrestrial ISPs.

As to latency; their tech woul;d be ~750[1] miles from the ground. They are targeting 1Gbps capabilities[2] vs a global avg of terrestrial companies around 20gbps. I don't know how viable that number is; but if it is physically possible then hitting 1/100th of that would still put immense pressure on ISPS as the floor for your weakest offering is 10gbps to be competitive. 2019 is scheduled date; so presumably 2020.

[1]https://www.engadget.com/2015/01/17/elon-musk-spacex-interne... [2]http://www.spaceflightinsider.com/organizations/space-explor...