Starlink is definitely cannibalizing the boutique WISP market, which for the most part probably deserved to be cannibalized, although it's sad for cool grassroots efforts like the parallel thread about Ayva. Many WISPs make a lot of their money on enterprise service, for example, construction trailers, rather than residential service. This type of service is being absolutely destroyed by Starlink, which while expensive, is much easier to install and maintain than WISP solutions, and tends to be much more reliable.
In the case of the linked article project, however, I suspect there's limited impact from Starlink. The linked article was a subsidy project. The subsidy is to provide fiber in places that don't have landline Internet options, so it's still eligible in spite of Starlink. With subsidies, Starlink isn't hard to compete with due to its price (customers would be rather irrational not to switch to a $55/mo service that was comparable to their $80-$120/mo Starlink).
Maybe you haven't had a crappy boutique WISP? They're genuinely awful usually.
My house isn't served by any landline provider (frustratingly, it's also _almost_ in a city, so there's no hope for subsidy access like the linked article), so over the years I've had a mix of corporate WISP (Rise), boutique WISP, and LTE as well as Starlink. I also work with a fairly large geographically distributed deployment of Starlinks professionally. Overall I would rate Starlink as excellent in the past ~9months or so.
It's been a pretty rapid improvement - a year ago I'd say Starlink was fairly, to use your phrasing, "dogshit," but now I'd say it's better than all competing wireless solutions I've used for moderate throughput needs (<100mbit) when the whole package is considered.
Good, well engineered fixed point to point ISM wireless can certainly be much faster and a decent amount lower latency but is a pain to install correctly, and licensed microwave is obviously superior at the massive overhead of, well, licensed.
On the other hand, Starlink is so easy - set the service address, plug it in, and a few minutes later it's Just Working with a nice /56 worth of IPv6 addresses.
In my experience WISPs usually take weeks or more to install, have terrible neteng, struggle with IPv6 and even public IPv4, deliver frequent backhaul outages, and often end up with bad enough buffer bloat to somehow deliver inferior latency to Starlink. For semi-mobile solutions like construction trailers, command centers, or even sensor telemetry use cases where LTE isn't available and you don't need a whole LoRa setup, Starlink is a no brainer compared to a WISP. For residential service it's certainly a last resort, but a superior last resort to all but the very best WISPs.
One thing we have noticed is that the plan prioritization is VERY aggressive in oversubscribed regions, though. The Mobile and Lite plans get quite aggressively chopped off compared to the Residential/Enterprise/Priority plans.
If they have half a clue regarding marketing and networking, they are doing fine. Starlink doesnt offer Layer 2 or Managed WAN options (Possibly Vocus is bringing these projects out at some stage on their behalf)
In dense areas, starlink underperforms. In larger cities Fibre is beloved. Theres a wedge, where WISPS are king and still are king, where the density is just right.
That said, if you are running a really shitty wisp, and you dont have any business links or complex services. And half your customer base just bailed for starlink, you will likely fold. But honestly, WISP as an industry can do without the cowboys.
However much you think it costs to lay new fiber, add three zeroes. If you lay a new line, you have to individually negotiate for easment rights with the owner of every single property your line crosses. You then have to pay for the construction and any remediations the property owner demands. Typically you have to put the land back to exactly how you found it.
This is the entire reason that the US forced Bell/AT&T to allow any company to lease a connection to their network. Without that, it would have been flatly impossible for any new provider to compete. There is simply no way that anyone could have ever built out a network to compete with Bell.
You can probably lease a connection to the fiber network in the same way (I haven't checked, but I assume common carrier applies), but if there's no fiber to the addresses you want to service you're SOL unless you want to front tens of thousands of dollars per customer. No ROI on that for many, many years.
Your only recourse there is to additionally stand up DSL/DOCSIS as a 'last mile' connection between the customer and your fiber. At additional unbelievable expense.
> If you lay a new line, you have to individually negotiate for easment rights with the owner of every single property your line crosses.
This may be true in some areas, but it wasn't in mine. I didn't have to agree to anything; it was negotiated with the town, not me. I presume they used the existing utility/sewer easements.
If I'm in a city neighborhood, could I just run fiber on the telephone poles just like Comcast does cable? I could probably run point-to-point connections from my garage to 16 single family homes and 2 multi-unit buildings with 3000' (extremely generous) of fiber.
If it was $50/mo, and 20 customers, that's only $1k/mo, which I'm not sure would cover a fiber backhaul...
You can get on the poles, if they're there and there's room, and you can find and follow the attachment rules and pay the attachment prices. Back when Google was going to run fiber to the home, they couldn't figure out how to manage the rules, which made their deployments very slow and eventually they gave up when AT&T (and others) deployed fiber to the communities Google announced before Google had managed to get plans finished.
They probably should have found a small telco or three to buy for expertise on pole bureaucracy.
Responding to your parent...
> You can probably lease a connection to the fiber network in the same way (I haven't checked, but I assume common carrier applies), but if there's no fiber to the addresses you want to service you're SOL unless you want to front tens of thousands of dollars per customer. No ROI on that for many, many years.
I would assume mandatory line sharing doesn't apply; the FCC walked back almost all of the 1996 Telecom Act line sharing; telcos and cablecos may well have designed their fiber to the home in ways to thwart what limited regulation was present anyway. If you're near 'commercial' fiber though, lots of that is available for lease.
Satellite is - once you've got the infra up in the air - very straightforward, with the downside that your satellite ISP is likely owned/operated by an unregulated billionaire nutcase that will turn off your access if he doesn't like you any more (c.f. Ukraine front line). It's hard to do that with regulated fiber backhaul, but not impossible.
I've seen a few wireless ISPs mentioned here before, which can be a nice hub/spoke model - run fast fiber to a community, but distribute via wireless (note, not WiFi) to homes and businesses within range.
I'd definitely love to see more community-run ISPs in the World, it's how the Internet should work, really.
> In 2022, Elon Musk denied a Ukrainian request to extend Starlink's coverage up to Russian-occupied Crimea during a counterattack on a Crimean port, from which Russia had been launching attacks against Ukrainian civilians; doing so would have violated US sanctions on Russia. This event was widely reported in 2023, erroneously characterizing it as Musk "turning off" Starlink coverage in Crimea.
Not true. From the very second paragraph in bold on top from your own link:
> Update: on 9 September 2023, Walter Isaacson said his biography’s claim about Starlink and Crimea was based on “mistaken” information [see footnote]
The footnote:
> This article was amended on 14 September 2023 to add an update to the subheading. As the Guardian reported on 12 September 2023, following the publication of this article, Walter Isaacson retracted the claim in his biography of Elon Musk that the SpaceX CEO had secretly told engineers to switch off Starlink coverage of the Crimean coast.
Starlink is definitely cannibalizing the boutique WISP market, which for the most part probably deserved to be cannibalized, although it's sad for cool grassroots efforts like the parallel thread about Ayva. Many WISPs make a lot of their money on enterprise service, for example, construction trailers, rather than residential service. This type of service is being absolutely destroyed by Starlink, which while expensive, is much easier to install and maintain than WISP solutions, and tends to be much more reliable.
In the case of the linked article project, however, I suspect there's limited impact from Starlink. The linked article was a subsidy project. The subsidy is to provide fiber in places that don't have landline Internet options, so it's still eligible in spite of Starlink. With subsidies, Starlink isn't hard to compete with due to its price (customers would be rather irrational not to switch to a $55/mo service that was comparable to their $80-$120/mo Starlink).
> and tends to be much more reliable.
maybe you don't have starlink?
the variance in latency is dogshit, especially in the evening when everyone in the area hops on netflix :) and i don't even play online games.
> maybe you don't have starlink?
Maybe you haven't had a crappy boutique WISP? They're genuinely awful usually.
My house isn't served by any landline provider (frustratingly, it's also _almost_ in a city, so there's no hope for subsidy access like the linked article), so over the years I've had a mix of corporate WISP (Rise), boutique WISP, and LTE as well as Starlink. I also work with a fairly large geographically distributed deployment of Starlinks professionally. Overall I would rate Starlink as excellent in the past ~9months or so.
It's been a pretty rapid improvement - a year ago I'd say Starlink was fairly, to use your phrasing, "dogshit," but now I'd say it's better than all competing wireless solutions I've used for moderate throughput needs (<100mbit) when the whole package is considered.
Good, well engineered fixed point to point ISM wireless can certainly be much faster and a decent amount lower latency but is a pain to install correctly, and licensed microwave is obviously superior at the massive overhead of, well, licensed.
On the other hand, Starlink is so easy - set the service address, plug it in, and a few minutes later it's Just Working with a nice /56 worth of IPv6 addresses.
In my experience WISPs usually take weeks or more to install, have terrible neteng, struggle with IPv6 and even public IPv4, deliver frequent backhaul outages, and often end up with bad enough buffer bloat to somehow deliver inferior latency to Starlink. For semi-mobile solutions like construction trailers, command centers, or even sensor telemetry use cases where LTE isn't available and you don't need a whole LoRa setup, Starlink is a no brainer compared to a WISP. For residential service it's certainly a last resort, but a superior last resort to all but the very best WISPs.
One thing we have noticed is that the plan prioritization is VERY aggressive in oversubscribed regions, though. The Mobile and Lite plans get quite aggressively chopped off compared to the Residential/Enterprise/Priority plans.
I run a small WISP - most of our new subscribers are coming from Starlink, but we are also cheaper and provide gigabit-class service.
Depends.
If they have half a clue regarding marketing and networking, they are doing fine. Starlink doesnt offer Layer 2 or Managed WAN options (Possibly Vocus is bringing these projects out at some stage on their behalf)
In dense areas, starlink underperforms. In larger cities Fibre is beloved. Theres a wedge, where WISPS are king and still are king, where the density is just right.
That said, if you are running a really shitty wisp, and you dont have any business links or complex services. And half your customer base just bailed for starlink, you will likely fold. But honestly, WISP as an industry can do without the cowboys.
This. How is local fiber not the easiest solution to the problem though?
However much you think it costs to lay new fiber, add three zeroes. If you lay a new line, you have to individually negotiate for easment rights with the owner of every single property your line crosses. You then have to pay for the construction and any remediations the property owner demands. Typically you have to put the land back to exactly how you found it.
This is the entire reason that the US forced Bell/AT&T to allow any company to lease a connection to their network. Without that, it would have been flatly impossible for any new provider to compete. There is simply no way that anyone could have ever built out a network to compete with Bell.
You can probably lease a connection to the fiber network in the same way (I haven't checked, but I assume common carrier applies), but if there's no fiber to the addresses you want to service you're SOL unless you want to front tens of thousands of dollars per customer. No ROI on that for many, many years.
Your only recourse there is to additionally stand up DSL/DOCSIS as a 'last mile' connection between the customer and your fiber. At additional unbelievable expense.
> If you lay a new line, you have to individually negotiate for easment rights with the owner of every single property your line crosses.
This may be true in some areas, but it wasn't in mine. I didn't have to agree to anything; it was negotiated with the town, not me. I presume they used the existing utility/sewer easements.
If I'm in a city neighborhood, could I just run fiber on the telephone poles just like Comcast does cable? I could probably run point-to-point connections from my garage to 16 single family homes and 2 multi-unit buildings with 3000' (extremely generous) of fiber.
If it was $50/mo, and 20 customers, that's only $1k/mo, which I'm not sure would cover a fiber backhaul...
You can get on the poles, if they're there and there's room, and you can find and follow the attachment rules and pay the attachment prices. Back when Google was going to run fiber to the home, they couldn't figure out how to manage the rules, which made their deployments very slow and eventually they gave up when AT&T (and others) deployed fiber to the communities Google announced before Google had managed to get plans finished.
They probably should have found a small telco or three to buy for expertise on pole bureaucracy.
Responding to your parent...
> You can probably lease a connection to the fiber network in the same way (I haven't checked, but I assume common carrier applies), but if there's no fiber to the addresses you want to service you're SOL unless you want to front tens of thousands of dollars per customer. No ROI on that for many, many years.
I would assume mandatory line sharing doesn't apply; the FCC walked back almost all of the 1996 Telecom Act line sharing; telcos and cablecos may well have designed their fiber to the home in ways to thwart what limited regulation was present anyway. If you're near 'commercial' fiber though, lots of that is available for lease.
Easements and access.
-The people that install fiber cost $100/hr+
-Trenching and directional boring equipment is expensive. Setting poles to run overhead fiber on is expensive.
-Restoration of green space, asphalt, and concrete are all expensive
-Easements are time consuming to negotiate if there isn’t a utility easement available to use
Laying fiber is expensive and time consuming.
Satellite is - once you've got the infra up in the air - very straightforward, with the downside that your satellite ISP is likely owned/operated by an unregulated billionaire nutcase that will turn off your access if he doesn't like you any more (c.f. Ukraine front line). It's hard to do that with regulated fiber backhaul, but not impossible.
I've seen a few wireless ISPs mentioned here before, which can be a nice hub/spoke model - run fast fiber to a community, but distribute via wireless (note, not WiFi) to homes and businesses within range.
I'd definitely love to see more community-run ISPs in the World, it's how the Internet should work, really.
And even if you aren't disconnected, there are valid concerns about privacy and censorship.
> owned/operated by an unregulated billionaire nutcase that will turn off your access if he doesn't like you any more (c.f. Ukraine front line)
Is this actually true?
https://en.wikipedia.org/wiki/Starlink_in_the_Russian-Ukrain...
> In 2022, Elon Musk denied a Ukrainian request to extend Starlink's coverage up to Russian-occupied Crimea during a counterattack on a Crimean port, from which Russia had been launching attacks against Ukrainian civilians; doing so would have violated US sanctions on Russia. This event was widely reported in 2023, erroneously characterizing it as Musk "turning off" Starlink coverage in Crimea.
Sad to see your comment downvoted for inconvenient facts correcting a very prevalent urban legend.
This is the level of political bias that exists where facts have to be suppressed to fuel the narrative and the agenda.
Original (non-primary) source cited at the end of that paragraph:
https://www.theguardian.com/technology/2023/sep/07/elon-musk...
[EDIT: Deleted quoted text from above, because I missed (!!!) the retraction from the cited primary author ]
Not true. From the very second paragraph in bold on top from your own link:
> Update: on 9 September 2023, Walter Isaacson said his biography’s claim about Starlink and Crimea was based on “mistaken” information [see footnote]
The footnote:
> This article was amended on 14 September 2023 to add an update to the subheading. As the Guardian reported on 12 September 2023, following the publication of this article, Walter Isaacson retracted the claim in his biography of Elon Musk that the SpaceX CEO had secretly told engineers to switch off Starlink coverage of the Crimean coast.
Absolutely. Apologies for missing that. I read the Guardian daily, and that makes it easy to miss the way they do apologies/retractions.