Google’s Equiano Cable Will Extend to the Remote Island of Saint Helena
spectrum.ieee.org
spectrum.ieee.org
Disclaimer: I work in two-way satellite telecom.
Until the advent of o3b, these were pretty much entirely dependent on various methods of connecting to the outside world via commercial geostationary transponder kHz leases, in the C and Ku bands. Works fine, but minimum 492ms latency return trip, and very expensive in monthly recurring cost per Mbps.
o3b has helped a great deal, in terms of monthly $/Mbps and latency. Further medium earth orbit and low earth orbit things (starlink, oneweb, amazon's kuiper) will help even more.
Due to the high cost of submarine cable construction, unless places happen to be roughly along the route that the cable would be taking anyways (as St. Helena is), there will remain a lot of places where 100% of the WAN link to the outside world will be some combination of LEO/MEO/GEO satellite access.
"Modern" satellite modems that are capable of advanced computation for low density parity codes and advanced FEC methods, plus adaptive coding and modulation to deal with Ku and Ka band rain fade, those certainly help a lot. It still costs a ridiculous amount to build and launch a 3000 to 6500 kilogram sized satellite into geostationary orbit. I am cautiously optimistic that SpaceX's lower launch costs due to reusable rockets will help with this.
Full deployment of the LEO starlink constellation economically seems to be predicated upon the use of a cargo-carrying reusable version of the Starship, which would be able to place many hundreds of satellites in one launch.
DfID have ruled out funding, and it’s unlikely St H will get any further EU regional funding.
So - either they’re going to be taking an almighty loan, or rely on the kindness of corporations.
I can’t see St H becoming a tech hub. It’s a lovely place, sure, but they’d need immigration of people with technical skills for that - I can count on one hand the number of saints with computer literacy, and they don’t want/can’t support any kind of immigrant population. Apart from anything else, all houses are occupied, and there’s precious little land upon which to build in Jamestown, and the saints do not want high rises and all the rest. No, I don’t speak for them, but I have spoken with them, and it’s pretty clear which way the wind blows.
I hope they do get a better connection to the world, but I fear practicality will get in the way. If they do plough ahead and assume the debt (I am assuming that there will be a debt to google), then the rumblings in Whitehall of making crown dependencies responsible for their own budgets might get louder. Their airport (which is perfectly fine, by the by, stunning approach) continues to attract flack as an ill conceived and expensive boondoggle.
But therein lies the problem. It’s unclear what the economic uplift would be, and how it would pay for itself. If it’s simply a convenient waystation for the cable, great, but it doesn’t seem to be, so to me, the reasoning is unclear.
The "SAEx" cable that's drawn into the map "failed to secure funding"...
Lagos, Nigeria
Accra, Ghana
Both have the advantage of being on the ocean (where cable landing stations are built), major population centres, major business centres, and IT/networking/telecom education and training programs are run in the English language.
If you are a mid sized ISP in Accra it is still quite costly to buy transport circuits going "both directions" - to get a 10 or 100GbE circuit from Accra to Cape Town, to peer with ISPs and content sources there, and also to buy a 10 or 100GbE circuit from Accra to western europe (Paris, London) to meet major IXes there.
From what I've seen looking at BGP topology for a lot of these ISPs, Accra is sort of a suburb of London.
Sad state of affairs really, because the costs are probably insurmountable for the islands inhabitants. I believe they are glad if they manage to maintain an airport.
Satellite based net access isn't my favorite, but probably the more realistic option.
Then I have news you'll appreciate. There's exactly such a place off the coast of Ireland. It's even actively recruiting for new residents.
https://people.com/travel/arranmore-tiny-island-ireland-look...
Huh. I'd expect vast bandwidth to be practically free if you're a tiny population living on the transoceanic cable.
I assume/guess the airport is subsidized by the British government.
At the moment. There are nine branching units built into the design, and only two are being used: Lagos and St. Helena. This leaves Google with flexibility for the future. Some possible options:
LOL, you mean "stringing another wire/fiber". I think we used to call this "bonding".
> Traditional subsea cables are powered from the shore end and rely on a dedicated set of pump lasers to amplify the optical signal for each fiber pair as data traverses the length of the cable. Now, SDM technology allows pump lasers and associated optical components to be shared among multiple fiber pairs, while still working within the unique power constraints of the ocean floor.
I don't profess to understand laser pumping, but it sounds like they found a way for multiple fiber pairs to share expensive and/or power-hungry active components. So it is stringing another fiber, but in layman's terms maybe kind of like network bonding by figuring out a trick to plug two ethernet cables into one port.
http://modegap.eu/?publication=space-division-multiplexing&w... (PDF)
Its a way of making QAM16 less crosstalky (well thats my reasoning. )
It's going to be "fun" for anyone who depends on something in orbit
You still have to go up, over, down, across wire, back up, over and down again
you're going to be paying a 25 to 50 ms latency on everything and that depends on having the dish tracking stuff working perfectly. (it is vastly better than the GEO stuff i've had to use in the past. but a replacement for fibre in general use. nope.)
oh and weather is still going to poop on your signal
"Why SpaceX is Making Starlink"
TL;DR -- satellites can be faster than cables in certain cases
Also marketing videos are always going to show best theoretical case. They are not going to show a cloudy day in a high contention network in a region not favoured by the orbit gods.
the speed of light in a mostly vacuum could (provided their meshing algorithms are up to it) make latency comparable but MEO sats already have a 125ms rtt time halving that distance isn't going to get rid of that penalty.
Sorry, was being an idiot. Latency does of course tend to mean a round trip, not just here to there.
Looking at the population of St. Helena I would be very surprised if the entire nation-state/island uses more than a single 100GbE wavelength back to the cable landing stations in the US/Europe.
From a landing station in St. Helana, access can be distributed by some combination of GPON fiber, terrestrial fixed PTP and PTMP wireless, maybe re-use of existing copper with vdsl2, g.fast, DOCSIS3/DOCSIS3.1, etc.
AT&T is lobbying the federal government for permission to put cell towers in Death Valley. Death Valley is fine without them. It doesn't need more tourist conveniences. It has"Death" in its name for a reason.
On the other hand, if this is something the people of Saint Helena voted for and approved, then it's up to them. I just hope it wasn't pushed on them by a bureaucrat.
Or at least, it'd probably increase rescue costs. And there are arguably better uses for resources than trying to make Death Valley a safe vacation venue.
There's a line.
You step over the line, you're responsible for yourself.
Driving to inhospitable and/or remote wilderness areas without preparation is over that line.
Plus, a sat beacon is $120. So people don't really have an excuse.
edit - doesn't mean putting up some radio cover is a bad idea, however I suspect it is not some panacea against badly planned road adventures. If you are capable of venturing forth into Death Valley without much preparation, you are more than capable of managing to forget a phone.
Expecting cellular coverage in places that are literally "the middle of fucking nowhere" is not reasonable. Remote parts of the western US states are too sparsely populated for this to make economic sense for cellular carriers.
People spend a thousand bucks on all sorts of frivolous stuff. For way less than that you can buy a gently used Iridium 9555 on eBay and a few new batteries for it, and a 20W folding solar panel.
My trip there in 2015: https://www.flickr.com/photos/defender90/albums/721576543876...