Slashdot Mirror
British ISPs Could 'Charge Per Device'
Barence writes "British ISPs could start charging customers depending on which device or which type of data they're using, according to a networks expert. 'The iPad created a very interesting situation for the operators, where the devices themselves generated additional loads for the networks,' said Owen Cole, technical director at F5 Networks. 'The operators said "If we have devices that are generating work for us, this gives us the ability to introduce a different billing model."' 'The operators launched special billing packages for it, which is in direct contravention to net neutrality,' said Owen. 'If things are left to just be driven by market economics, we could end up with people paying for the amount of data that they consume to every device and that would not be a fair way to approach the market.' Owen also foresees a billing system that charges less for non-urgent data, with an email costing less per bit than either Skype or video packets that need immediate delivery."
Congrads, you got first post. But was the Urgent Packet Delivery Fee worth it?
OK, not really, but it is really fucking stupid.
I would advise against this type of "hypothetical model" unless you want to slow innovation and business growth.
I would also advise against it because the industry is leading consumers into an "online world", where all data will exist.
If infrastructure can not handle the load (how much dark fiber do we have in the world?), then it needs to catch up. Living off the 90s infrastructure boon is just not going to cut it.
...and now our bandwidth too? When will this madness end?
93rd rule of Slashdot: No matter how obvious my sarcasm is, my comment will be taken seriously by someone.
"could" "might" "maybe", what a complete non story.
broadband ISPs COULD charge you per character typed but they don't and probably wont.
"It's very expensive maintaining and upgrading network equipment. We're due for another upgrade soon. Wouldn't it be nice if we didn't have to upgrade, but could still put our prices up? Can anyone invent a reason for this?"
Who will get the business?
None of this drivel seems to be coming from ISPs - just a technical director at F5 Networks.
Since so much stuff seems to come in over tcp/80 nowadays I'd like to see how they propose to reliably differentiate between HTML pages and images, *Tube videos and downloads of device firmware updates, Linux .isos, etc. - or are they just going to charge based on the size of each request? <1MB at 1c/MB, <10MB at 2c/MB, <100MB at 3c/MB, >=100MB @ 10c/MB? Why have monthly caps at all then?
These people seem like simple leeches to me. You just want an internet connection. Your probably connecting to your own router doing your own networking.
That's one connection
So you give me the internet and I'll give you the cash. Nobody needs to get screwed.
Wait... Your company bribed a politician, didn't it.
How long will it take before someone mods DD-WRT to obfuscate Internet traffic to make device identification by ISPs difficult?
Some dude says ISPs could do random shit including charging people whatever they think they can get away with. He also uses unclear and awkward wording*.
Film at 11?
*Like where he says "'If things are left to just be driven by market economics, we could end up with people paying for the amount of data that they consume to every device and that would not be a fair way to approach the market."; I think he means the exact opposite of what he seems to say?
Everyone knows fat people use it more than skinny people, why not start charging by users weight?
What if someone used all their devices through a single router?
2 years ago I got an Android phone on my own (not through my Operator). I called them to add 'data' to my plan and they wanted to know if it was an iPhone or an Android as they had 2 different plans. They were the same price so I investigated a bit. It turns out that they block http requests if the referrer field doesn't contain 'Android'. Like that's gonna stop me from using the phone as a 3G hotspot for the rest of the bus, right.
Non-Linux Penguins ?
First they wanted us to buy our music repeatedly
I'm pretty sure "per second playback billing" is next on RIAA's list.
I means these blokes are in boardrooms licking their proverbial chops, and we are on the pick wheel.
Its look like the rapacious beginnings of the cable industry all over again, but this time you count amongst you shaledowns fees for your refrigerator's call to the repairman. 'wonder if there will be an opt out for that?
its looking spooky, people.
Umm. If I'm using an iPad then I'm probably not using my desktop or laptop and creating load there. Why not charge based on the number of people in the household, that would make more sense. Or gasp, charge for the amount of bandwidth used. But if you start breaking the billing down like that then people probably shouldn't have to pay a fixed monthly fee any more, but we can't do that how will ISPs make tons of money!
http://www.openbsd.org/faq/pf/scrub.html
Additional speaker charges are after that. Whoa you want stereo? That'll cost you twice as much.
lol. How you do pay 5.1 x 99c?
$5.05 (rounded up of course) + a low frequency surcharge of $0.50
If it's a stereo to 5.1 upmix, you don't. You pay 16x.
Unfair, I hear you say? But no! You've got your left channel, your right channel, your center (using data from left and right channels), your left surround (using data from left and right channels), and your right surround (using data from left and right channels).
Clearly that's eight separate audio channels in simultaneous use, requiring eight times the licensing fees. And you do have two ears, right? So you're listening to each of those eight channels twice over...
Now, pay up, serf!
Translation:
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
The fucking greedy cunts
Remember where the internet comes from ? Look at what greedy people have made of it, what a bunch of assholes...
I can see where they are coming from, in a sense: you should pay for how much you use, which is hard to argue against. After all, that's how we pay for other resources we use - I don't use the internet for watching movies or other high-bandwith things, so why should I pay more to support those that do?
However, what they propose is almost exactly the opposite of paying for what you use; it's like being billed for water by measuring the size of your garden or the number of taps in the home. And just as for water, it is perfectly easy to measure the actual consumption; if they don't know how, I am sure there is a large proportion of /. readers who can help them figure it out.
"with an email costing less per bit than either Skype or video packets that need immediate delivery" ..!! god save those poor ppl..
with this analogy it might seem they will charge us for delivering the e-mails instantly.. !! or charge the email based on the distance they need to be sent..!! a postal delivery model suits the bill.. videos and skype can be treated as freight
if anyone wants to limit free speech and put obstacles in the way of internet use, leave it to the Orwellian UK government, as the British people sleepwalk into even more darkness.
iPads don't use anymore bandwidth than any other device will that you can watch over the air video on. iPads cannot in principle do anything at all any other computer cannot do. This is pure gouging. Note that it is the cellular carriers themselves that have pushed video on command. The goal is good enough broadband that these and many many other applications can run for everyone everywhere. This is not achieved by nickel and dime-ing us.
Keep it down, will ya? AT&T might hear this!
Daily read for tech news: Freezenet.ca
I can see where they are coming from, in a sense: you should pay for how much you use, which is hard to argue against.
Let's say the cost of running an ISP transmitting x bytes is ax + b for some constants a and b. How large is a relative to b? I think a lot of the cost of running an ISP is in infrastructure and wages, especially for support. What's the resources usage when transmitting a packet one hop? The electricity to run the router and the space occupied by the router. How much is that, one nanobuck (given that packets come in rather often)?
Okay, so maybe I'm pulling figures out my ass. But I think it would be interesting to look at the books of an ISP to see what the costs are. I don't think they are measured per byte.
There is a big difference between water consumption and Internet consumption. With water you're depleting a resource, and whenever you use it or however you use it, the amount you consume is the amount it's depleted by, so that's how much you pay for. With Internet you're not depleting anything -- the links are still there with the same capacity, after you've gone.
Instead, on the Internet, what you need to be charged for is the "hurt" you cause others by your usage. If you use 4Mb/s at peak hours you're causing lots of hurt, if you use 4Mb/s in the middle of the night you're not causing much hurt. Or if you download 100MB at 1kB/s you're not causing much hurt but it's for a long time, whereas if you download it at 10Mb/s you're causing a lot of hurt for a short time. How it all balances out is rather tricky to understand. Arguably, time-of-day throttling as a crude attempt to approximate this idea of "hurt".
Thats perfectly fine. I will use a proxy and then I will only be technically using a single device from your point of view! But hey since you cannot actually tell what devices I have in my house that use the internet without digging into my data they I will be using ipsec to somewhere else. Of course though you have to get all isp's in the UK to change to this billing model together otherwise all your customers are going to leave and join the other isp. This is also fine because the crap isp's that are coming up with this stuff don't work anyway.
Why is the iPad costing them more work? The article refers to it as the "traffic hungry iPad". Traffic hungry? A PC downloading Torrents every day is not traffic hungry?
There's some really garbled understanding of what is going on there.
What I think is fair is something along the lines of the following:
1. Pay some fixed cost per unit time in order to have a connection.
2. Pay per bit sent and received based on QoS.
It seems like the most fair thing to me. Uncapped is just rediculous and a complete lie. The companies shouldn't even be allowed to claim it since it is blatantly false advertising.
Part 2 is the most sensible option. People pay a reasonable price for what they use. Of course it only works if they charge a non punative price per bit. If ISPs want to offer some automatic capping to prevent enexpected bills too, then that's fine too.
It also avoids any network neutrailty problems. If you want low-latency, you must pay since it costs more to implement. If you want to run your bittorrent client with VOIP QoS, then fine. Knock yourself out.
Remember, QoS is not in violation of network neutrailty if it is selected by the user. If the ISP offers only uniform QoS to the user, but then nobbles companies that don't pay the protection racket, then that is very much in violation of network neutrality.
SJW n. One who posts facts.
Any reputable engineer who isn't owned by one side or the other in this 'debate' will look at the network infrastructure, then the size of the anticipated customer base (hell, just for Apple's projected sales alone), and the anticipated customer usage patterns. Result is a train-wreck. No other result. It won't work.
Now I'm an unusual customer with normally unusual demand and, fortunately, all my wireless service provider does after a I blow through twice the max capacity for the month in just a couple of days and just slows my connection. The rest of the industry either cuts you off or charges you exorbitant overage fees. If everyone wants video wherever, whenever (or downloads a lot of alpha and beta software to test), it just won't work.
Engineers and economists (usually) deal with the real world, the world with (rational?) constraints. I am, and have been, both to my misfortune. Why misfortune? Because I've been watching this build for a very long time. No one listened. Enjoy.
"[I]t is a wise man who admits the limits of his knowledge or skill, and that pretending either causes harm." --Terry Go
So, the technical director of a large carrier-grade router and packet classification equipment manufacturer is suggesting that British ISPs adopt a billing model which requires carrier-grade router and packet classification equipment to operate?
I'm not sure that an article should really be allowed to claim that something is an opinion of "experts" but quote only one (admittedly expert) person whose business would directly benefit from his prediction being accurate. I'd rather they actually asked an academic or someone else without direct economic interest (as well, not instead).
I think this article was more aimed at the ISPs going to the meeting than the rest of us: "Hey British ISPs, if you want to be able to charge more than just £x/megabyte, how about this model? We also happen to be able to sell you the equipment to implement it. You probably should get the government to agree first, if you happen to be meeting with them any time today."
Paul "TBBle" Hampson
Paul.Hampson@Pobox.Com
"Uncapped is just rediculous and a complete lie. The companies shouldn't even be allowed to claim it since it is blatantly false advertising."
Well, I on FIOS I have a 50Mb/s connection, and I probably download 500GB per month, and have done so for about 5 years.
If I put my mind to it, I'm sure I could download more, and I don't think Verizon cares.
Is it rampant speculation week on Slashdot? First the ridiculous "Apple's handcuffing web apps!" nonsense from the Reg, and now this completely speculative nonsense? /. standards are really slipping. Can we link to some proper journalism please?
Yes, I must be new here.
It's a Unix system - I know this.
You also have the freedom to prostrate yourself on a prayer mat facing some arbitrary direction five times a day. And you have the freedom to get as angry as you want about your detractors, reprimanding them for discounting "any possible scenario where inventing a supernatural being might be useful".
Unfortunately, to engage anyone other than the choir in discussion you have to provide a supporting argument for your position. A market deals with scarcity, both by restricting use (some sense of right to control property) and by innovating more efficient uses. This is why people have meagre allocations of IPv4 addresses and why IPv4 NAT exists. The same does not apply when you have 128 bit addresses.
British Telecom has legendary reputation for being an asshole ISP in all regards, shapes and service...
i know quite well how hard they suck... and we don't even have BT in germany... that's how angry BT customers are... imagine.. :D
Can't you just use a proxy to get around the lack of NAT?
Firstly, the extra volume created for ISPs by iPads is close to zero: they're being used as extra devices in houses, and aren't capable of running any of the bandwidth-intensive P2P applications that (when they're pimping different things) ISPs and vendors are keen to tell us represent 90% of their volume.
Secondly, this is a vendor of DPI kit pushing applications for DPI. But it's a doomed endeavour. It would be impossible to split tariffing based on numbers of devices as the market would react with domestic proxies if NAT didn't provide enough aggregation. So the only way it could conceivably be done would be by inspecting packets at close quarters to see which application is being run. At which point the market would respond with encryption.
It does apply if they try to charge you per IP. I'd sure as hell NAT my devices then. Try reading the context to his post. I'd think it would also be useful if you have any still useful IPv4-only devices at home, an IPv6 NAT could enable that device to interface with the outside world by doing IPv6 DNS resolution, etc for it.
which is totally what she said
Now you want to turn the volume up? There's some more potential for other people to hear it! That'll be ( $100 ^ increase in decibels) thanks.
which is totally what she said
You just demonstrated the Anon's point perfectly. mirix gave a reason for users to want to NAT IPv6 - to avoid per-IP billing. You then say a lot of hoopla without addressing the point that IPv6 NAT would be useful in a per-IP billing situation.
Is per-IP billing stupid and unwarranted with IPv6? Yep. Will it exist? Almost certainly.
totally nor fair for us in the UK
imagine per cost billing for ringers...
Also, bandwidth that is not used is wasted...
Water that is not used can be stored and used later.
http://spamdecoy.net - free throwaway anonymous email - avoid spam!
UK ISP's announcement about voluntary commitment to net neutrality?
As the island of our knowledge grows, so does the shore of our ignorance.
Anon is trying to use an extreme exception to prove a rule. IPv6 NAT would also be "useful" in the case where a government implemented a law requiring everyone to deploy IPv6 NAT just because, or in the case where a weekly lottery was held only for those people deploying IPv6 NAT. But it is intellectually dishonest to justify a rant that IPv6 NAT is therefore sometimes useful.
IPv6 involves giving each network a /64 and there are enough of those for everyone. Conscious effort is required to do otherwise. If some ISP makes the effort to create an artificial scarcity by limiting routing of IPv6 IP addresses then it might as well just stick with IPv4 and IPv4 NAT. The main advantage of IPv6 is the opportunity for end-to-end connectivity.
Anything can be argued useful if given the right context. So you have to limit the definition of "useful" to reasonable (technically, socially) scenarios. Finally, if you're not the type to bend over and take anything coming, encountering an absurd re-definition of "useful" should be taken as an opportunity to voice a loud objection and refuse to participate in the nonsense. Consider: Airport body scanners are useful. Random stop-and-search laws are useful. Censorship is useful.
I'm pretty sure "per second playback billing" is next on RIAA's list.
Pay-per-play for musical recordings has been around since the 1890s. See "Jukebox" on Wikipedia.
Somebody mod the parent up, he posted as Anonymous.
Amazing how the NATsi's modded him to -1 for bringing up the idea. Just goes to prove his point.
I'm not a lawyer, but I play one on the Internet. Blog
it forces the users devices to either continually cycle over for hours on end trying to send traffic
Is that such a bad thing? If a home user wants to torrent while sleeping, he could buy a cheap little low-power ARM NAS and use it as a home seedbox.
...finding more ways to charge you more for the same service you've had for years.
Hey ISPs? I've got a mind blowing idea, how about you ACTUALLY IMPROVE YOUR SERVICE to keep up with today's standards, instead of trying to live by the standards of the 90s.
What do I know, I'm just an idiot, right?
What a load of rubbish. If an ISP started doing this, they wouldn't have any customers left.
ISPs that lack have to consider such crazy ideas, probably lack the infrastructure to support their existing customers.
Certainly, implementing this idea will eliminate the majority of their existing customer base, and solve that problem.
Between phones and different computers I currently have 10 different items off my internet connection, how would the ISP know that? Are they running IPv6 with each device using their own public IP? Are they going to be using their own routers locked to specific mac address and preventing you from buying your own?
Does my ISP know when I watch a youtube video that it is going to a desktop or a laptop or an iPad? It doesn't seem like they should be able to tell that.
just wait for comcast to do this with ipv6 $5+ per system just like how in some areas they want $8.95 per cable box and $16+ per HD DRV.
I would NAT regardless, just to guard against any incoming ports that might have been enabled by default.
Anthropic principle: We see the universe the way it is because if it were different we would not be here to see it.
as i understand it, at the very highest levels, the debits & credits for a connection is based on how many more bits one backbone generates vs. another backbone over each others networks during a billing cycle.
Further consideration, at the consumer level, this is why home users ABOVE the 99th percentile are the ones targeted for either cessation of service or increased charge.
every day http://en.wikipedia.org/wiki/Special:Random
> Owen also foresees a billing system that charges less for non-urgent data
That is frigging AWESOME! I can't wait to wire into the mind-reading system that will tell the ISP which data is urgent. Particularly when I'm running data through an encrypted tunnel.
It's also going to have to make a very good estimate of the difference between my concept of urgent and that of every other user on the same shared channel. That will be an extraordinary advance in real-time psycho-analytics.
Unless they are talking about letting me choose when to run my line at low-latency -- which would actually be pretty cool.
Stop-Prism.org: Opt Out of Surveillance
I mean if everything is run behind a router (though I guess you might need to add your own) how would they know how many devices are being used?
Look, this is not about whether we have an exact analogy between water use and internet use; otherwise we would be back in the "tubes" scenario, right? I'm just saying, it is hard to argue against paying in proportion to how you use, be it bandwidth or not.
It isn't all that difficult to find a reasonable model - here, meaning one that most people would find agreable, rather than "the most objectively fair" (whatever that means). Assume there is something like a price per minute on the total bandwidth on the internet, leaving out local variations etc since this is only a sketch anyway. So, you pay for the percentage of the bandwidth you use per minute, measured by counting the number of packets with your name on per minute as a percentage of the total capacity. This is not even mildly difficult; it is trivial.
It will be like what I have today, one device connected to their network a router.
Net neutrality implies there is no preferential bandwidth.
IT DOES NOT MEAN YOU DON'T HAVE TO PAY FOR WHAT YOU USE.
I remember Comcast (and probably ISPs in general) doing this way back when broadband was first being rolled out. You were supposed to pay per computer hooked up, and have a modem for each PC. I recall having to put my router away when the cable guy came to fix problems because you weren't supposed to be doing that.
Why would you have to NAT in order to accomplish that? Any port based firewall can do this without the overhead, and busted ass stupidness, of NAT.
I live in Scotland, and I recently signed up with an ISP that did this. They have block control over my building, so there wasn't any other solution. They have a device (I'm pretty sure it's just a linux box based on my nmapping) that looks at each packet's TTL. If the TTL is odd and the port is NOT 80 or 443, it drops the packet. If the TTL if odd and the port is 80 or 443, then it redirects you to a billing page. I bypassed it by incrementing TTL at my gateway. I imagine people will modify openwrt/dd-wrt to do this as well. Additionally, I have a solution which tunnels my connection over a VPN to an Amazon EC2 instance and does some magic to beat QoS. It seems like oversubscribing at at least 2000% seems typical. I'm paying for a 50 mbit/sec connection. I see closer to ~6-8 mbit/sec (no, I'm not getting my megabytes and bits confused). Additionally, since I know a little about my ISP internally, they run all these blocks to their HQ over MetroE-like products, and then concentrate it into their core. Their transit is 10GigE (based on traffic numbers pulled from various private sources). I know that they sell far more than 10GigE of bandwidth to customers. There is far more wrong with the ISP environment here than you'd know. My girlfriend's ISP is even worse. They do a significant amount of throttling on specific ports (1935, 80, so on..) and basically any kind of download during the day slows down to a halt. They have about 6000 people + several offices on a 10 GigE connection. I don't know if this will even be seen, but I thought it was worth throwing out there.
Local power companies are changing their cost structure as well. They are creating a 3 tier cost model. Tier 1 is power being supplied to your heating and cooling (water and air), Tier 2 is power supplied to any permanently wired lighting, and finally tier3 which is all standard power outlets in the house. The power company believes that the different tiers have vastly different demands and requirements and to ensure they can maintain power when it is needed, they will charge more for the T1 and less for the T3. Some T3 devices can also be moved up to the T2 and T1 level depending on their function. If a portable heater is plugged into an outlet it might actually fall under the T1 requirements. Same with a portable lamp moving to T2.
A power company exec stated that they used to just "supply power" at a specific rate but now with the tiered model, we can charge according to the demand and still maintain a consistent experience. "Home heating and cooling units really need a lot of power and it is critical that homeowners get it. We can not guarantee they will get it unless this tier cost structure are in place, it is fair."
You should use a firewall as a firewall not NAT. You are woefully uneducated.
...my wireless router, so bill me for one device.
No, I'm actually just thrifty. A hardware firewall costs much more than a simple NAT router.
Anthropic principle: We see the universe the way it is because if it were different we would not be here to see it.
expect more jailbreaks in the future...
Unfortunately, to engage anyone other than the choir in discussion you have to provide a supporting argument for your position.
Let's consider the possibility that I wish to disguise my internal network topography (eg. make it appear that I have only one device, as might be useful in the context of the British ISP's) while also allowing for inbound connections to be served by various machines via a single, unchanging nexus IP. That is to say, what if I want the privacy features of a forward proxy in conjunction with the abstraction layer of a reverse proxy? Unless I am mistaken, simultaneously satisfying both of these goals would preclude the use of IPv6 address privacy extensions. Furthermore, I imagine that ISPs could decide to issue a single static address per account and then block all other traffic from one's network (I believe this would also preclude the use of privacy extensions).
I have nothing against IPv6. I was just pointing out that there are useful scenarios for NAT. Note that I deliberately didn't cite a 6-to-4 NAT or the use of the scheme as an ersatz "firewall", because I didn't want to engage in a freewheeling migration or security discussion. For our purposes, let's assume the world is a happy place: v6-only and everyone has as many SPI firewalls as they want/need. Furthermore, please note that I never insinuated NAT was without drawbacks.
All I said (above and now), is that the "NAT == evil" zealots seem unable to acknowledge any scenario where NAT might be useful. I believe some such scenarios exist, and even as a lurker reading the IPv6 NAT discussion threads on various forums I found the tone of the discussions to be... well, like your response.
Have you asked yourself why you got angry instead of chuckling at my original post?
Anon is trying to use an extreme exception to prove a rule. IPv6 NAT would also be "useful" in the case where a government implemented a law requiring everyone to deploy IPv6 NAT just because, or in the case where a weekly lottery was held only for those people deploying IPv6 NAT. But it is intellectually dishonest to justify a rant that IPv6 NAT is therefore sometimes useful.
Whoa, that appears to be a strawman that you just beat down magnificently.
One practical, useful, and real context for IPv6 NAT is the exact scenario under discussion: allowing one's network topography to be disguised and to be presented to the outside world as a single host via forward and/or reverse proxy. We are not discussing artificial shortages of a virtual "resource" anymore, we are discussing privacy and the ability to masquerade.
Believe me, I understand the dream of what end-to-end publicly routable addresses might do and how it would eliminate whole classes of hacks (ie. I can admit NAT has serious drawbacks). Why does it seem that you cannot acknowledge that NAT might have some benefits for certain scenarios?
Finally, if you're not the type to bend over and take anything coming, encountering an absurd re-definition of "useful" should be taken as an opportunity to voice a loud objection and refuse to participate in the nonsense.
In the US, most people have at most, say, two or three choices for broadband ISP. I have found that most of them implement virtually identical policies, so it really isn't "bending over and take anything coming" when no provider offers you what you desire and you have no ability to negotiate their policies.
Besides, they are sneakier than that. If you protest, they will say, "Oh, we offer that, too! Our business broadband connectivity is only $249/month and offers all those features you desire!"
PS. You can avoid airport body scanning if you charter a flight. Don't just take it lying down, vote with your dollars! Feel free to pay $5,000 per flight instead of $300. Yes, it's disgusting, but it's also the sordid reality we share.
No it does not. In fact many of NAT routers can and will act as a hardware firewall when you disable the NAT.
I'm pretty sure "per second playback billing" is next on RIAA's list.
rounded up to the next whole hour, of course.
You can make a fine hardware firewall out of an Asus RT-N10 which costs about $25 at the store. How much cheaper do you want?
"Patriotism is your conviction that this country is superior to all other countries because you were born in it." -- GBS
This rant might make sense if you completely ignore the context of the discussion, which is about how IPv6 would make it easy for ISPs to see how many different devices people were using and charge accordingly.
Nobody is trying to take things outside of that context except for you, and you are seemingly only doing it for the purpose of justifying a rant.
"Patriotism is your conviction that this country is superior to all other countries because you were born in it." -- GBS
Gaming the system will lead to retaliation by ISPs in the form of DPI, throttling, and other nasty tricks. My point is that it's worth thinking hard about how pricing should be designed, not simply going with a version that "most people would find agreable" and that will end up surrounded by kludges. There is in fact an IETF working group called conex, working on how to measure "how you use it" in an un-gameable way. This should be a sound basis for un-gameable pricing. More reading here.
It was my understanding that the zealots prevailed, and IPv6 NAT was declared a "nonfeature".
As if nobody is going to make software that does it anyway.
What is a business supposed to do when their ISP gives them IPv6 pubic addresses but they still have thousands of IPv4 computers with private IPv4 addresses and site local software that doesn't support IPv6?
So you have to limit the definition of "useful" to reasonable (technically, socially) scenarios.
So you want a real use for IPv6 NAT? Information security. If I have several devices and I don't want the outside world to know how many devices I have, or be able to tell which is which, I can use NAT to make them all appear as one public IP.
Yes, most of the reason for having NAT is not present with IPv6. No, that doesn't mean it shouldn't be possible. It just means you probably don't need it most of the time.
If I understand the idea correctly...
it would be like the power company charging you separately for EACH device you've plugged into the wall. Moreover, rates would be dependent on WHAT the device was, not how many WATTs it uses. You enjoy your TV more than your 500watt toaster? the TV costs more. 3 ipods drawing 5 watts each will cost more than that 1500watt spaceheater...
Ahhhhh...the sound of silence! Pleasant. Birds chirping, dogs barking is pretty music to my ears. Who really needs the crap RIAA is selling? Go find some free indie music if you can't live without; but try without first - you might be surprised.
Cable providers here in the bay area, ca, usa used to charge more if they detected more than one MAC on your cable modem (hub). They also prevented known router mfgr MACs from obtaining IP addresses - luckily this was avoided by configuring your router's MAC to spoof the MAC of your original machine 's NIC (back then it was called "@home" for me)
from 09 F9 11 02 9D 74 E3 5B D8 41 56 C5 63 56 88 C0
to 45 2F 6E 40 3C DF 10 71 4E 41 DF AA 25 7D 31 3F
Give me a non NAT static IP for every device, and I will pay for each device.
Violence is the last refuge of the incompetent. -- Isaac Asimov
You're the third person not to know about IPv6 Privacy Extensions.
ISP's, including mine, routinely advertise "datarate" for "price". The price must be paid in full every month, accurate to the penny. The datarate? Is it ever even close to what is advertised? The advertised datarate is always higher than what is provided. They claim that prices need to be high to 'build out' the system, but when they build out, they provide more services, occupying all of the new extra bandwidth --- for a higher price! Sure you paid for their new infrastructure, but you will never see any of it. They also claim that more competition is bad.. but what that really means is they aren't making a killing. So now British ISP's are complaining that people are actually using the net they are already paying for, and so they see the need to charge more. One of two conclusions: 1) litigation against ISP's or 2) competition in the market will force them to not be so greedy. The only other scenerio is that the government does something American like granting a monopoly and limiting competition, in which case, the locals will doubly suffer: 1) locals get gouged for internet services, and 2) internet based companies move to where they can get on the net for less.
You're the third person not to know about IPv6 Privacy Extensions.
Yes, and IPv6 Privacy Extensions attempt to obscure the topology by generating more IP addresses.
That's exactly the opposite of the goal in the context under discussion: the ISP's charging per-IP. The goal is to have one's network topography appear to be a single host with an unchanging IP, the better to hide from ISPs that charge per IP (and block all traffic to any IP they haven't whitelisted in their gateway).
IPv6 Privacy Extensions are a lame attempt to do what NAT does without NAT. And it doesn't even work -- if you have five PCs each with one IPv6 addresses all connected to the same host at the same time, it's obvious that you have at least five PCs. Moreover, if different machines have different usage profiles then you can track them individually as they change their addresses based on their usage profiles, instead of having all usage aggregated behind one IP address. And making machines change their addresses with a higher frequency can actually make it worse because it makes it more likely that a machine will change its address in the middle of a TCP connection, which will have to be reopened using the new address, making it pretty obvious what happened.
To make it work fully you would have to assign multiple IP addresses to each machine simultaneously, one for each connection it has open -- but that's just NAT by another name, using part of the IP address in place of the port number.
This is, alas, quite reasonable. The point of NAT being evil with IPv6 is that it breaks many assumptions of IP/TCP/UDP. Meanwhile, since with IPv6 your ISP is supposed to assign you a subnet block large enough to accommodate millions of devices (this is why IPv6 has such a large address space), you can avoid the breakage of NAT because you've got enough addresses. Using more of that block that is assigned to you doesn't cost the ISP anything extra.
In other words, NAT is evil since there is no legitimate technical reason for it. Whereas this is a political issue.
Fortunately, there are no ways to detect whether a NAT router is being used.
Wait, no, that's trivial.
And there's no active research going as far as trying to count the number of devices behind a consumer NAT router for the specific purpose of stopping people from exceeding the AUP on connected devices.
Oh, never mind.
Like I said, IPv6 NAT has no uses.
That doesn't negate the benefit of NAT in general, it just means the existing implementation is flawed. So we want to have the IPv6 NAT router rewrite the packets' IP ID and TTL to avoid identification, etc.
It's not "flawed" for this reason. NAT's purpose is not to hide the count of hosts behind the NAT gateway and it's never done this effectively. But it does seem fairly typical of NAT proponents to have a gravely unfounded sense of security when using it.
If you want to engage in an arms race with your ISP to hide the count of hosts behind your network, you're welcome to do so. There's enough DPI already going on at ISPs that you're wasting your time to think you can win the race with nothing but a consumer gateway. IOW it would not even be sufficient to create a hypothetical perfect NAT implementation (heh) which avoided the "etc." that no-one's exhaustively enumerated.
The discussion started off concerning a flawed method of breaking the ISP AUP, but Anthony Mouse suggested the benefit of a false sense of security. NAT solves neither problem (see my other posts in this thread).
What will inevitably be the case is confusion. Like now with phones and mobile phones people barely know what their bills mean or if the option they have is the cheapest for them according to all the bit of this and that deals available. Thus will be the case with provision of the internet until some genius comes up with a simplified service where you charge one fee and just use it, how radical would that be (in 3 or 4 years time I mean:-).
You're still arguing against the implementation rather than the concept. Security is always an arms race -- if someone finds a vulnerability then you have to plug it. The fact remains that if you want to hide the number of hosts behind a firewall, step one toward achieving that is to make them all appear to have the same IP address.
Dude, vigorous handwaving is no substitute for actually confronting the facts. I'll try it one more time because you may just have a genuine misconception, but I'm AFK after this...
There is no such concept as "1 IP address one machine" (as NAT itself demonstrates!) so you are making a conceptual error if you think that one of NAT's purposes is to hide the count of machines on your network from your ISP. Your ISP has every packet you send and receive available for a decent analysis if it really wanted a machine count. It could at the very least trivially confirm that your network configuration is designed to give the wrong impression of how many machines you're using.
To reiterate, the flaw is not in the implementation; it's in your conception of what NAT is.
Now, NAT may help obscure the particular machine used on a network from a remote host, but IPv6 privacy extensions do a better job of this (I can choose where and when they're applied). An application level proxy may be an even better solution in some cases, as it does not simply pass on a subset of identifying qualities of the original machine. In every case you're hoping that the remote host makes a sufficiently crude analysis of behaviour or other fingerprint.
There is no such concept as "1 IP address one machine" (as NAT itself demonstrates!) so you are making a conceptual error if you think that one of NAT's purposes is to hide the count of machines on your network from your ISP.
I'm not sure what you mean by "purposes" -- things can be used, or modified to be used, for purposes other than those they were originally designed. And some things are effective for purposes they were never designed for -- NAT was effective in the 1990s at preventing ISPs from knowing how many devices were behind a dial up modem, because the methods for detecting were either not known or not feasible then. If you want to continue using it for obscuring the number of machines going forward, it has to be changed to defeat the vulnerabilities we now know exist in using it for that purpose.
You seem to be taking the position that those vulnerabilities cannot be removed. For example:
Your ISP has every packet you send and receive available for a decent analysis if it really wanted a machine count. It could at the very least trivially confirm that your network configuration is designed to give the wrong impression of how many machines you're using.
How can they do this, in a way that is impossible or even impractical to work around? It should be possible for a NAT router to emit packets from two machines that are byte-for-byte identical to the packets that would be emitted if you ran exactly the same programs at the same time on a single machine. Current NAT implementations are not designed for this obviously, but you seem to be arguing that it is impossible to achieve.
I understand that it may be impossible under specific circumstances. For example, if you have 5000 machines, it may be impossible to make them appear to be one machine, if only because of the traffic level. But I have confidence that you can make e.g. two machines, one running a web browser and the other a BT client, appear to be a single machine running both.
Those brave enough to voice positions contrary to the "NAT == evil" party line will receive vituperations along the lines of "you are thinking about it wrong" or "no one needs that".
I also forgot to mention that when pressed to the extreme, they (much like the Git jihadis) will also eventually resort to insinuations like "that never worked/existed, anyway, so it is no loss to be rid of it." Then again, I really didn't expect such a textbook demonstration by the zealots to an obscure, offtopic reply to someone's idle musing about potentially NAT'ing IPv6.
NAT obscures the topology of the network just fine. Yes, I even read the links that were supplied. Yes, there are theoretical attacks, but as you pointed out this will turn into a cat & mouse game. Anyway, you can tell an argument is vapid when it dismisses the present utility of a system based on the possibility of the development/deployment of future countermeasures.
"Dude, what's your malfunction?! SSL/TLS is worthless! Asymmetric Key Public Key Cryptography can be broken by quantum computers (*mumble* if they spend decades & tons of money to research them), so that's why you are wrong to want support for SSL/TLS in the next gen socket layer in Linux! Smarter people than you have already discussed this and the decision has been made. <they then insert random links here about Shor's algorithm in an incoherent attempt to support their argument>"
Heh.
NAT solves neither problem (see my other posts in this thread).
I read them.
NAT may not yet be a perfect solution to the masquerading issue (due to the theoretical attacks you cited & possibility of cat/mouse with ISPs); however, NAT seems to have much more potential to help solve the issue than the dogmatic IPv6 approach, which appears to involve sticking one's fingers in one's ears and humming "there are no valid uses for NAT!" over and over.
So, given a scenario where ISPs charge by the IP and only allow traffic to/from whitelisted IPs, what useful suggestions do you propose to evade the ISP AUP and masquerade as a single host? Furthermore, this should not require additional cost (eg. no private VPN to Sweden, though that would probably be the gold standard). Current NAT works fine for these purposes, and it would rapidly evolve (eg. TTL masking) if ISPs attempted to start ferreting it out.
If you think there isn't a demand for this, you are mistaken. If you think router manufacturers wouldn't go along with this, you are mistaken (cf. the prevalence of current v4 NAT'ing routers' ability to MAC clone to "help" connect to restrictive ISPs). If you think people wouldn't jump through hoops and/or put up with "busted ass stupidness" to avoid paying more money, you are mistaken (cf. the wide prevalence of P2P piracy since the late 90's).
As others have pointed out, someone is going to make v6 NAT, if only for these reasons. If people are willing to spend hours poring over an oscilloscope/LA to break hardware encryption on consoles, there are certainly many more people who are willing to tack a relatively simple "bag on the side" of IPv6 to evade ISP AUP's or to mask their topology as a single host for the lulz. Furthermore, they won't care how offensive it is to anyone's technical aesthetic sensibilities.
By holding to a dogmatic refusal to consider NAT, it is likely that the eventual system that is deployed will be less refined than it could be, and if it achieves a critical mass then it will just be something that network admins will have to live with (fait accompli).
Wishing this away won't help; besides, if we are wishing then I want a nuclear submarine or a moon rocket.
When you buy a beer at a bar, they charge you the same price whether it's the top of the barrel or the bottom, the first bottle in the case or the last. There's no sacred reason that internet usage should be "all you can eat", since ISPs are in business to make money. I agree with you that "congestion pricing" would be better to balance out usage; and then they offer "quality of service" choices, and suddenly we wind up with a tiered service level controlled by those who can pay, just like space on supermarket shelves where small or new players can't even get a spot. What I'd like to see, ever since grade school: Punish the troublemakers. Don't punish everyone, don't raise everyone's rates, just find the IP address that's maxing out its bandwidth all day and deal with its user. Heck, maybe he'll be happy to find out that his system is pwned.