FCC questions ISPs' selective memory about data caps America's network regulator wants to "better understand" why ISPs still cap consumers' data usage even though they need more and providers have shown they have the "technical ability to offer unlimited data plans." As the Federal Communications Commission (FCC) pointed out, many American broadband ISPs – along with their …
I can understand data caps on mobile data contracts, but data caps on (residential/domestic) broadband? I'm in the UK and the last time I had a data cap was about 15+ years ago on a low cost contract where the peak-hours daytime service was capped at about 150GB/month (no cap on overnight usage). Nowadays, I'm on an unlimited gigabit FTTP connection which has a list price of £29.99/month (and quite a bit less with various discounts/deals available). Even the low-cost social tariffs in the UK have unlimited usage. In fact, that's an important part of those contracts - fixed price, no surprises.
So how on earth is it that there are widespread data caps on broadband contracts in the USA? - ineffective regulation? a lack of actual competition in the marketplace? cartel behaviour from broadband providers? Without uniform access to low-cost high-speed connections, the economy will suffer. Yes, broadband providers might do well in the short-term, but long-term there will be lower revenues/profit.
I pay $90/mo USD for 250 down, 10 up. That's with a 1.5 TB data cap. If I ever have to redownload my steam game collection to a new computer I'm going to go well over that, and if I backup my GOG collection it'll probably be the same. Not to mention any regular internet use and video streaming. The problem is that there really is only one broadband option in my area; when internet was initially built out, Telco lobbyists managed to convince local legislators that they needed a monopoly in order to recoup costs. Unfortunately that monopoly is still in place, which means I get subpar internet for premium internet prices.
My ISP is 'gracious' enough to permit me to go over the data cap one month out of each year without charging the extra, which simply means if I do go over, I straight up saturate my internet as much as possible for whatever benefit I can get as long as I don't go over again that year.
I pay that for 930 down. 940 up on fiber with no firm limit on data versus the 250 down, 10 up that I had 11 months ago. Nice to have more than one broadband option. Also very helpful to live in a relatively affluent tech heavy community where fiber operators stand a reasonable chance to get a return on their investment installing fiber.
Going back 25 years, the local cable company was offering 3Mbps down with a 2.5GB monthly cap.
The president of Concast acknowledged that data caps are not about network performance. They're about revenue. Greed is the only reason they exist. Well, that and many of the broadband networks in the US are legacy cable TV systems, bludgeoned into acting as data networks. They do this...poorly, judged by both economic and performance metrics. Fiber only systems tend not to have caps, and their speeds are higher and at lower cost, but they are newer to the market and not as available. Also interesting...where there is fiber/coax competition, data caps tend not to exist. Comcast tried to put them in after COVID in Massachusetts, but the state regulators told them to shove off with that "stuff", so we still don't have them here. We also have a fairly robust fiber market (which runs faster and cheaper than Comcast)
US Network availability
Fiber 10,000 Mbps 39%
Cable 1,200 Mbps 89%
DSL 140 Mbps 88% <--- this max speed is a bit of a laugh...you need to be in the CO to get it
Fixed wireless 100 Mbps 46%
So how on earth is it that there are widespread data caps on broadband contracts in the USA?
A combination of many things. So cost, complexity, marketing, lobbying and tradition. Oh, and regulation, which is lobbying by other means.
But why would you accept data caps on mobile, yet not broadband? Especially when that's frequently sold as 'mobile broadband' and a wireline replacement? Techincally and commercially, that's really the same issue, ie managing capacity on a finite resource in a FRAND kind way, so Fair, Reasonable and Non-Discrimatory.
Spectrum is finite on both fixed and mobile, so can easily become congested, especially where you've got lots of subscribers. Services are marketed based on theoretical capacity, ie xMbps but with finite capacity on the backhaul, so can easily become congested. Then we end up with complaints about being sold 'Up to xMbps', but getting much less. Solving that problem means throwing capacity at the problem, but that increases costs. If costs increase, we lose the social benefits, unless those connections are subsidised to reflect that benefit.
The alternative is to manage capacity more effectively, ie using QoS/CoS to try and prioritise time sensitive traffic when congestion occurs. Congestion will almost always occur, because TCP/IP is just very bad at dealing with that because it assumes the clever stuff happens at the transmission/transport layers.. And it's made worse by software, ie relying on UDP, which is extremely dumb vs TCP, which is slightly less dumb. For business users, they can run VPNs that use a combiination of DiffServe and MPLS, for consumers, well, it's too bad.
Main challenge to managing resources efficiently is a combination of political and lobbying, so it's basically the heart of the 'Net Neutrality debate. Regulators could solve some of that by setting the rules, but there's a lot of lobbying to keep the Internet best-efforts.
Bandwidth is always finite, BUT with mobile connections, the load is random and variable. I have no doubt the mobile carriers [are] taking unfair advantage of their customers, but I will concede that a cell tower can serve 100 people on a random Thursday, only to be hammered on Friday when the nearby stadium fills for a Taylor Swift concert. With home internet, the number of users can be budgeted and accommodated, so anything less IS a matter of choice by the ISP- not capacity constraints. Mobile data caps MAY be a money-grab, but residential data caps ARE....100%
Bandwidth is always finite, BUT with mobile connections, the load is random and variable. I have no doubt the mobile carriers [are] taking unfair advantage of their customers, but I will concede that a cell tower can serve 100 people on a random Thursday, only to be hammered on Friday when the nearby stadium fills for a Taylor Swift concert.
It's still the same issue as with fixed line, ie effective network and capacity planning. You know there's a stadium or other venue in the area, so you know you have to plan for that peak traffic. Plus with 4 & 5G, attempt to off-load as much trafffic onto unsuspecting WiFi operators to make it their problem. And there are attemps to take advantage of their partners, ie the contracts for stuff like mobile offload aren't exactly generous. The Internet is free though, isn't it? If you're ever in the position to do one of those jobs, try suggesting a usage based charge priced based on the mobile operator's data tariffs. They don't seem to think the idea of a fair revenue share is fair.
With home internet, the number of users can be budgeted and accommodated, so anything less IS a matter of choice by the ISP- not capacity constraints. Mobile data caps MAY be a money-grab, but residential data caps ARE....100%
Nope, it's pretty much the same issue, other than mobile being perhaps more physically constrained by available spectrum & non-interference. Provisioning a 1Gbps PON (Passive Optical Network) for a typical FTTC/FTTP (Fibre To The Curb/Premise) is technically easy at the access level. So fibre per subscriber converted into a standard wavelength via WDM makes it easy to run 10x subscribers as 10x 1Gbps(ish) onto a 10Gbps(ish) wavelength and drop it into something with more brains.. And more cost. So from the cab/headend switch where it becomes Ethernet and maybe MPLS or IP to the Internet, you start loading costs on. Uncontended, that gets very expensive, very quickly, simply due to the cost of anything intelligent enough to make a decision based on a packet. 100Gbps interfaces aren't cheap, and also take up a lot of space & power.
So while it's 'technically feasible' to offer an uncontended service, it's astronomically expensive.. Especially when you can't use Huawei any more. Plus it's inefficient* because you know that for 99.9% of the time, your users aren't going to be pushing 1Gbps 24x7x365 so you contend the network. Then you find a 'heavy user' who is trying to do 1Gbps all the time, so craft a way to discourage people from hogging all the capacity.
And then you hit the elephant in the room and the reason behind the millions spent on 'Net Neutrality lobbying. The heavy user may not be aware of the traffic they're generating because maybe they've downloaded a P2P app, and their connection has become part of the app provider's distribution network. Or just the long running streaming debate. Streaming generates a lot of traffic, but streaming providers contribute very little towards the cost of running the access networks. So the age-old peering and settlement issue. Absent any fair cost sharing, the costs fall entirely on the access networks, and then their customers. So the only way to fund capacity increases is to charge their users, which increases connection costs.
Usage charges are currently a 'fair' way to charge based on consumption/resources used, and until the regulatory model changes. Personally I'm a fan of shifting to a muni-model where access is treated as many other utilities. Whoever controls the roads controls the wayleaves, so run fibre, maintain that and sell access to any content provider that wants to try and exploit it.
Downvote for understanding the tech, but not understanding business.
If you oversubscribe a link based on your assumptions of what an 'average user' will consume, that's your business model. If you're 'wrong' because of the heavy user, well that's because you failed to accurately model usage upstream. We've been at this long enough that it's not a novel thing, so this argument doesn't hold weight.
Also you throw out cost-based arguments without backing up your assertions. A Cisco C9300X-24Y costs around $25K today. That's a 24-port switch, so my slice of it is 1/24th of that. So roughly 1K of that. Then there's aggregation-layer switches and core switches above that, but my 'slice' of the traffic cost becomes a smaller and smaller fraction even as the costs go up. Let's call the hardware cost of my hypothetical residental fiber link in a brand new development $5k. Those switches (and based on my experience of ISP practices this is being generous) will be in services for about 10 years before being upgraded. I know, the true picture is messier than that as the whole network won't just be forklifted all at the same time, but from a high-level view, it's a practical assertion.
That puts the ISP's hardware costs, to recoup their 5K investment, at 120 months, at about $41/month to give me a 1GB link without caps for 10 years. I pay $90/month for half that, with traffic shaping/data caps on top of that. I know there were fiber build-out costs, and they were expensive, but that's a one-time investment with decades of payoff. it doesn't affect the long-term math that much, or nearly enough to make it somehow necessary to do throttling/capping except for greed purposes.
Edit to add: I went with Cisco in my hypothetical scenario. I doubt ISPs were using top-shelf companies for their switches. As you imply, they were probably using cheap brands before, so their costs are even smaller than my scenario.
This is not about the tech, it's about greed. What's actually happening is they're not building out enough capacity, and holding onto ancient hardware much longer than that to squeeze dollars out of customers. That 10 year switch? it's more like a 20 year switch. They recoup the costs on year 10, and just pocket all the rest with no upgrades for the next ten years, and when they can't keep up, they institute caps to squeeze people into the older hardware, longer, so they can continue to pocket the difference.
Also you throw out cost-based arguments without backing up your assertions. A Cisco C9300X-24Y costs around $25K today. That's a 24-port switch, so my slice of it is 1/24th of that. So roughly 1K of that
Except your switch is dumb, and just (mostly) suitable for aggregation. This is intentional on Cisco's part because even though it's switches used to be able to route (ie the good'ol Cats could be persuaded to run IOS), that functionality was disabled. However, cheaper switches are available that don't have Cisco's margins or management overheads slapped on..
That puts the ISP's hardware costs, to recoup their 5K investment, at 120 months, at about $41/month to give me a 1GB link without caps for 10 years. I pay $90/month for half that, with traffic shaping/data caps on top of that.
That ignores many costs, like taxes (UK still has rateable values applied to 'lit' fibre), staff costs, wayleaves, cost of capital.. And as I said in my post, it only gets you as far as a 24-port switch. One of the reasons sane ISPs don't use Cisco (other than cost, management complexity) is they're office kit. They don't like being plonked inside a cheap street cabinet, so costs are increased in either providing an 'office in a box', or locating the switch in a PoP somewhere with a more stable environment. Which then means renting a rack in an existing exchange, or leasing/buying/building a PoP somewhere convenient.
Then you completely skip over any of the backhaul and other cost components. So you've filled a switch in mid-Wales. Now, you just have to plumb that into the Internet (well, your AS). So you have 24Gbps of uncontended capacity into.. what? You could contend and lease a 10Gbps circuit towards London. But then you need (or should need) diversity, so find that. So no problem, just lease 6x 10Gbps.. Except your 'cheap' Cisco* 24Y doesn't have an enough ports. So no problem, you could use a 25Gbps uplink, except then you need to find a supplier of 25Gbps for your WAN. Or someone who'll lease you dark fibre, but then you need to also find/build regen sites every 40-80km or so.
But at some point, you're also going to need a router to do the IP thing. Those are really expensive, especially when you're needing multiple 100Gbps interfaces. They're also pretty power hungry, need cooling, need someone to poke them with a stick and stop them crashing etc. Plus you'll need some kind of solution for mail, DNS, oh.. and peering or transit. Oh, and staff who can answer that age old question "Is the Internet down?". Oh, and maybe a lawyer, because your customer is now running VoIP, the Internet is down, and their house is on fire..
So there are a lot of costs involved, and it's still back to the same issue. Absent any other funding sources, the only people who pay those costs are their subscribers. It's a strange business given it's pretty unique in it's funding model, ie ISP pays all the delivery costs. Most households have a similar service delivered over wires, and are 'happy' to pay a standing charge + usage charge. The telephony world sorted out (mostly) a setllement system for origination & termination cost sharing. The postal system has been charging for delivery based on usage for far longer..
*WTF is '48 mGig'? milli-Gigabits? And 'ThousandEyes Enabled'? Surely you only need to support Five.. It's a shame Cisco switched from being a tech company to a marketing machine.
You seem to have missed that and talk about a specific product choice I pulled that particular switch from a 2 minute search to get ballpark pricing on a high-end access-layer switch.
The _point_ was to pick something way overpriced for what you need for an access switch, for uncontended bandwidth (which you don't need for residential internet at the access layer) and trivially show that even if you WAY overengineer in this way, they're still overcharging and don't need data caps.
I'm not sure you read or understood what I wrote or its purpose, but I'm not here to argue details with you. When you can use back-of-the-napkin math and five minutes of searching to see that a company's claim is BS, you don't need to dive into bills of material or company financial sheets.
I also gave a nod in my reply that there were other costs, but it should have been clear based on order-of-magnitude (charging 20x what they need to cover equipment outlay) that data caps aren't a technological necessity - they are a profit feature.
Because radio is somewhat more limited than running cables.
5G is fast bit there are limits. If you have dozens of people going flat out on 5G connections, you quickly run out of space for more users and crucially, shit starts getting unstable...when you start nearing a certain usage level on 5G it all goes to shit pretty fast.
I worked on a project to build a live streaming platform that broadcast live video feeds over 5G. The goal was to hit around 12 simultaneous 1080p feeds, which is theoretically very easy on 5G when you do the maths based on bandwidth alone (you can in theory go up to 64 QAM and have 64 feeds)...but in practice we were able to achieve 8x 720p feeds without any hitches or 3 1080p streams.
The observation was that 5G is very good at bursty stuff, but rubbish at constant throughput.
We were in a perfect world situation as well. Signals all went over cables.
To be fair though, everything we did was in software because the hardware to do what we wanted doesn't exist...or didn't at the time.
5G is fast bit there are limits. If you have dozens of people going flat out on 5G connections, you quickly run out of space for more users and crucially, shit starts getting unstable...when you start nearing a certain usage level on 5G it all goes to shit pretty fast.
Yep, same as the Internet. ISPs generally want to avoid congestion because then bad things happen, ie dumb apps just blasting out packets, and if they think they're going nowhere, blasting out some more! Then routing sessions drop, everything re-routes, and repeat until flap dampeners kick in.. But how quickly people forget-
https://datatracker.ietf.org/doc/html/rfc791
1.2. Scope
The internet protocol is specifically limited in scope to provide the functions necessary to deliver a package of bits (an internet datagram) from a source to a destination over an interconnected system of networks. There are no mechanisms to augment end-to-end data reliability, flow control, sequencing, or other services commonly found in host-to-host protocols. The internet protocol can capitalize on the services of its supporting networks to provide various types and qualities of service.
.. except where prohibited by law. Which is one of the frustrating things about the way the 'Net Neutrality debate has been politicised by vested interests. Postel was a smart chap, his customer was the DoD and he knew the weaknesses of IP, and the demands of the customer. Hence why the IP header has had ToS/QoS bits from the outset, it's just we're not allowed to use them to provide better quality of service..
Where that has been made illegal, it is because ISPs don't intend to use those to improve the quality of service, but to extract more money from their customers. They do this effectively by tricking them: yes, we will give you unlimited bandwidth, with speed only restricted for congestion, unless it's somebody who has paid us, in which case their competitors will run slowly. The way they explain this is "video constrained except from certain providers". That has two effects: in the short term, ISPs can try to negotiate extra payments from users to reverse the decision to break their service, and in the long term, they can try to sweat their infrastructure for a bit longer. One effect it doesn't have is improving the quality of service.
The way they explain this is "video constrained except from certain providers". That has two effects: in the short term, ISPs can try to negotiate extra payments from users to reverse the decision to break their service, and in the long term, they can try to sweat their infrastructure for a bit longer. One effect it doesn't have is improving the quality of service.
Nope. QoS/CoS really only works when there's congestion, and selectively degrades traffic based on assigned priority. So stuff like 'WRED', or 'Weighted Random Early Discard' randomly throws packets away out of it's buffers. That can be bad because often the app will just resend a whole chunk of traffic and create more congestion. IP is just carp, as the original RFC explained. Other policies can prioritise traffic so it's sent first. This can be a GoodThing(tm) and is done.. usually quietly. So if you're a regulated carrier announcing a new service, you point out every circuit has a 128Kbps prioritised chunk. Because all wholesale customers get the same service, it's fair and non-discriminatory. Then the regulator may helpfully suggest that this is suitable for VoIP, because that's a safety-of-life services and they may fine the bollocks off you if people start dying again, as they did with early VoIP offer.
But it doesn't 'break the users service'. It might in fact make it better. Plus the 'user' generally has no service with the ISP, which is the problem. They create congestion, but don't pay the ISP to deliver any of the traffic served. They don't contribute anything towards the cost of capacity upgrades needed to deliver 4K adverts, even when they make lots of money pushing them to the ISP's customers, who generally don't want them. It's that disconnect in billing relationship that creates the problem, ie if content won't pay, the only people to bill for upgrades are the ISP's customers. Regulators may decide to ban usage caps or charges, but all that will happen then is service quality degrades due to increased congestion, and ISPs have even less revenue to fix that via upgrades.
The QoS features built into hardware are usually intended to be implemented when the network is congested. The ones that net neutrality advocates are arguing against are not. Even net neutrality adherents accept contracts that, for example, allow users to be deprioritized when the network is busy based on their total usage, and they complain instead about ISPs trying to be selective about which traffic they will affect. I have already explained what the ISPs have done and would like to continue and expand which lead them to this opinion. I have a few other points about your last comment though.
"the 'user' generally has no service with the ISP, which is the problem. They create congestion, but don't pay the ISP to deliver any of the traffic served."
At first, I thought you were talking about the customer of the ISP. From your next sentences, this appears not to be true, but in case you were including both of them, the customer is certainly paying plenty to the ISP for all their stuff. But, as I say, this sentence suggests you were talking about the other end:
"They don't contribute anything towards the cost of capacity upgrades needed to deliver 4K adverts, even when they make lots of money pushing them to the ISP's customers, who generally don't want them."
Which is how the internet has generally worked. The ISP concerned is sending that data on because their customer has payed them to do so, and they received it from the ISP from the service provider, which has likewise been payed to do so. Those two ISPs may negotiate an agreement where they exchange data for free or where one of them gets money as a result, but either way, they will figure out a way to deliver the data from one paying customer to another one. That the data sent makes money for someone is irrelevant; they have already agreed to provide the service by signing a contract with the customer, and it is their responsibility to do so no matter who is making money using the service or how successful they are. Similarly, if I use an ISP-provided connection for something which makes me no money, I don't get to either have my connection for free or send the bill to every site I used, because I looked for a company that agreed to exchange bits and payed one to do so. Your electricity provider will not charge you more if you're using it to charge your work computer, and thus, get money. Your water provider will not charge you more if you're using that water in baking something that you sell, thus turning a profit. Your ISP doesn't get to do it either.
The answer, as it usually is, is follow the money. Data caps are lucrative, and the caps are set so as to get most people to go over. That includes the ISP refreshing data, pulling telemetry, ect. That is one thing I like about my service, no caps. I can run my max data rate 24x7 if I want. It's only a 10MB connection but so far it more than meets my needs.
Data caps are lucrative, and the caps are set so as to get most people to go over. That includes the ISP refreshing data, pulling telemetry, ect.
Nope, they're actually a PITA. If you have a capped service, you need something that can measure usage, pass that to a billing engine, produce a variable bill and have staff to handle the complaints. That's a cost. It's much easier and cheaper to just produce a flat-rate bill every month. The ISP also probably refreshes very little data, or telemetry because again, that's a PITA for generally no real benefit.
The far greater problem is all the apps and junk people have running on their machines (ie Win11) that GB updates because their developers are too lazy to do incremental updates. Oh, and then all the 'telemetry' those apps also generate from watching your every keystroke.. And then of course there's adverts. If those actually cost something to deliver, we might get fewer of them.
If there was a *need* for data caps, why wouldn't ISPs implement them like cellular providers? When you hit your cap, your internet speed just slows down. About 20 years ago, my first DOCSIS cable modem was set by the ISP to do more-or-less this. You can pay for faster speeds for the rest of the month if needed, but no surprise bills. And this cap MUST be prominent in all the advertising, of course.
Or perhaps ISPs want to offer a pay-as-you-go service for those who choose. Light users will have $2/month bills.
The combination of the two... you MUST sign-up for an all-you-can eat service, but if you eat more than we think you should we'll quietly penalize you, is completely abusive and indefensible.
Data has a cost.
What we pay for is burst speeds. That moment we want to download a large game without needing to wait. Or have a film quickly fill the buffer so it starts playing immediately.
We want low latency.
So most people use reasonably small amounts of data. The problem comes with you have a user who consumes vast amounts, maybe switches between two 500GB games, unable to store them on the device so deletes and downloads time and time again.
That user on an unlimited connection can do that. But they’re consuming a disproportionate amount of network resources. They could be making others experiences worse and increasing the provider’s infrastructure costs causing everyone’s bills to increase.
So caps are a safety net also. A way to manage a shared network so everyone gets a reliable and predictable service. Predictable is what business users need.
Data has a cost. What we pay for is burst speeds.
Quite the opposite. Large ISPs are Tier-1 or Tier-2 providers. They do NOT pay for data transfer at all. Instead they have zero-cost peering with a lot of other networks. Maintaining burst speeds during peak usage hours is in fact the expensive part, as that's where they may decide they need to upgrade their equipment, but once they have the peak capacity sorted out, it's nearly zero cost to operate as much data as you want to send through it. And that's not just theory, there are public filings from the large ISPs where they put it on the record that removing data usage caps will have negligible costs to their network operations, and they promote the data cap fees as pure profit.
You probably have a home network. How much more does it cost *you* when you're transferring large files between two systems on your network, versus when everything is idle?
If ISPs wanted to throttle their heaviest users during peak periods, I doubt anyone would mind. But they don't because that doesn't earn them any extra money.
Or alternatively, the USA joins most of the rest of the industrialised world in not having data caps on fixed line accounts. Which would definitely be better for consumers.
Most UK broadband accounts these days are unlimited but with a "fair usage policy". This is broadly drafted but is only ever likely to hit consumers who are downloading huge amounts of data during peak hours. I've never been clipped by a fair usage clause despite a household with three users, many devices, much streaming, and my job as a developer and consultant either doing a lot of Zoom or pushing GByte sized images to container repositories. You really have to go nuts to provoke bandwidth limitations due to the fair usage clause.
I'm US with unlimited but "fair use" policy. I'm not positive, but I think I've hit that policy every month, which (per the ISP) sets my traffic to lowest priority. Streaming is probably the biggest consumer at my household, but I also do cloud-based backups (Backblaze). Of course, I'm not positive if, or when, I'm being backburnered...
Take a look at what is actually generating external traffic on your home network. Forced advertisements. Chatty IoT devices. Devices that generating traffic when not in use - such as bandwidth coming from a TV when it is on HDMI1. Tracking telemetry for everything.
Block list DNS lookups with Pi-hole. Put advertisement blockers in browsers. Disconnect IoT devices from the Internet when they are not needed. Of course this only applies to us techie people, the great unwashed masses have no clue.
It really is a racket to charge customers for bandwidth usage, and then pack that bandwidth with undesired content.
It's a question of supply/demand/competition.
In a capitalistic market, any entity is going to try to charge as much as possible for as little as possible, as that will enable it to charge more for a little more, etc...
If you're an ISP, even if doshing out 10TB costs you (practically) the same as doshing out 1TB, but you believe that the few that will need more than 1TB will maybe accept to pay more, then you're not going to give for free something that you can charge for (your shareholders would not be happy if you give away something that you can charge for).
Giving more or less at a given price will depend on what the competition is doing also of course.
Then there's a small detail that can really peeve off an ISP. Consider this:
-You decide to pay for an unlimited no-cap multi-Gbit fiber link at your typical residential monthly price of (for example) $50/month, and then (as you're the techie type), you decide to cable up all your local residential unit (50 of them...) and share the link with them.
Apart from flagging your excessive usage (typically the 50x residential hanging off one single box is going to show up...), and then going after you, its hard on the ISP to be able to pinpoint what you're doing, and shut you down (that sharing of 50x residentials, if it catches on, would just kill the business of the ISP).
During the pandemic, I can understand that as people were all at home, a lot were hitting caps if those caps were too low and it was probably more hassle to try and enforce them (and take a PR hit) than the "good" publicity gained from appearing to lend a hand in hard times.
Apart from flagging your excessive usage (typically the 50x residential hanging off one single box is going to show up...), and then going after you, its hard on the ISP to be able to pinpoint what you're doing, and shut you down (that sharing of 50x residentials, if it catches on, would just kill the business of the ISP).
It's actually pretty trivial. Just look at the number of sessions going through the device. Then refer the subscriber to the 'not for resale' clause, and perhaps let Ofcom know someone's running a public service without an operator's licence.
(Also try this sometime. Typical home routers can cost <$10 in bulk and can have hardware limits on the number of simultaneous sessions, especially if they're also NAT'ng. Those can also sometimes be configured as limits. Used to be a fairly common TT when 'smart' customers decided a consumer/home connection was much cheaper and should just work supporting an office with 200 PCs. YGWYPF..)
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
