Hacker News Reader

Comment by seatac76 | Hacker News Reader

This feels like a capability that will have giant impact across geopolitics which is not immediately clear.

They mentioned backhaul, I wonder what is the upper bound of scale, will it be big enough to complement/replace submarine cables at some point.

View on HN

Nah submarine cables have tons and tons of capacity. Fibre is super efficient and you can even use multiple colors to add capacity (DWDM). Radio spectrum is much more limited. Also they don't just lay one fibre at a time. Each fibre has its own clean broad spectrum where radio is always shared.

Sats will not replace them for backhaul. The capacity isn't there. For limited emergencies sure but not for all our Netflix.

This is all great for users in poorly connected areas and that's what it's really for.

Ps: also the sat link is more for remote areas, the network is much more efficient when an uplink is available in the same area and this is also how it provides near cable-like latency. Basically a bent-pipe system. If you'd use this capability for long-distance back haul you'd need to tie up a lot of sats' capacity increasing latency and decreasing capacity overall.

Satellite service is more likely to replace cell. If the latency gets low enough, and capacity gets high enough, we can expect the Starlink/T-Mobile partnership to expand beyond emergency services. I'd certainly trade speed and latency for total planetary coverage.

More likely cell will replace satellite, as it is cheaper and more reliable.

No it's not.. It's far most cost effective to deploy cellular. Fiber in the ground to a tower will easily last 50 years. The radio hardware will need ongoing maintenance, but it's still far cheaper than launching 30k satellites into space with a lifespan of five years.

Are users in poorly connected areas enough to be profitable? These satellites have a short lifespan so it'll have to be hella profitable to make up for the constant replenishment launches no?

no, that's why they're losing money on starlink and keep raising prices

Well for reference: "The current lit capacity of the Southern Cross Cable network is upwards of 5.4 Tb/s, while the Hawaiki cable between NZ and Hawaii is 10 Tb/s. Together with upgrades planned soon, this comes awfully close to Starlink’s worldwide system capacity, quoted to be 23.7 Tb/s worldwide or somewhere between 15 and 20 Gb/s per satellite. Except that Australia and New Zealand between them have a population of just over 30 million — and a lot of CDNs with a local presence. But there are billions of unconnected and under-connected users out there waiting for service." https://blog.apnic.net/2021/06/11/leo-satellites-part-4-why-...

Yeah, the new satellites being announced here have about 100Gbps apiece. They want a constellation of 30,000, have been approved for around 7000 so far IIRC, so they’d have 700Tbps total (less than that usefully as most will be over water). But they plan on another big upgrade by about a factor of 3 with Starship, which should bring aggregate capacity to 2Pbps. This will be most useful for places like Hawaii as the satellites will otherwise be idle, so will have lots of space capacity.

There are also competitors like OneWeb (smaller, but still useful) and Kuiper (not yet running, but potentially a comparably sized project, a project of Amazon’s), so this is just getting started.

This is a really poor way to compare. even with cross-links about 30% of their bandwidth is useable in the best case. starlink has no way of increasing capacity to their heaviest demand areas besides launching thousands more satellites. fiber is much, much, much more efficient.

I consider about 20% of the aggregate bandwidth to be usable (the land area vs total area of the Earth), but in special cases like Hawaii, there’s potential a huge amount of bandwidth since Hawaiians can have access to some of this 80% otherwise-unused portion.

Sure, but the problem is that the places you might have high bandwidth might already have good access to internet. Hawaii in general has very good internet already, so asking them to pay a premium for satellite internet isn't always a good trade.

That’s not necessarily the case. Hawaii has limited bandwidth and higher latency.

Starship is vaporware.

Let’s place a bet on it. By the end of 2025, there will have been more successful Starship orbital flights than SLS orbital flights.

Total international Internet capacity is 997 Tbps, according to Google. If they need more, though, they’ll presumably add more wavelengths to existing fiber. That puts Starlink at about 2%.

Reply on news.ycombinator.com