Withdrawl of IPv6 Technical Trial

[#1]

If you have two PCs at home can you and your partner simultaneously log in to different accounts on the same bank without any issues? We've done that with our bank. As that works successfully over NAT on the home network I can't see any reason why there should be any issues extending that to CG-NAT. If there are any problems it will be the bank's fault not the ISP's.

[#2]

It's hard to imagine a practical failure scenario whereby two NATed streams could somehow overlap once established.
Furthermore, given that nearly all banking access is performed via HTTPS the crypto keys for each session would not match hence even if the the network layer somehow swapped things round the data exchanged at the higher level would effectively be gibberish.

[#3]

Isn't the REAL problem with IPv4, is that there are many legacy products (for example the Xbox 360) which don't yet support IPv6.
The suppliers of these products will say it's not worth updating as customers aren't demanding IPv6 as their ISP doesn't provide it.
Because of these legacy products, ISP customers will continue to need to provide IPv4, irrespective of the IPv6 roll-out.
Rather than implementing CG-NAT to extend the life of IPv4, surely the answer is for all the ISPs to rapidly provide IPv6.
If the majority of ISPs supplied IPv6, there would be renewed interest in broadband products, customers would start using IPv6 on their iPhones and other gadgets,  but crucially customers would be hammering on legacy product manufacturers support helplines DEMANDING a firmware update for their Xbox or Roku video streamer, etc.
The final trick, would be for the clever ISPs to introduce a worthwhile monthly subscription DISCOUNT for broadband customers who use a connection that is IPv6 ONLY, reducing the pressure on the IPv4 pool.

At the moment the pressure is heaped against the ISPs, firstly because IPv4 address space has virtually gone, and the increasing demand for IPv6 provision.
If ISPs provided IPv6 NOW, then the pressure would be somewhat shifted onto the internet connected device manufacturers to provide IPv6 connectivity, because there would be the demand to do so.  It would also pressure websites that don't support IPv6 to go dual-stacked.
If an discount for IPv6 existed, then there would be an incentive for customers to take an active interest in their ISPs product and a sales opportunity to upgrade from legacy to current products.  The IPv6 discount would have the effect of 'crowd sourcing' the pressure on manufacturers for firmware upgrades.

If this happened fast enough, then roll-out of CG-NAT might be avoided.  As it stands it all feels too little too late 

[#4]

Quote from: purleigh

Isn't the REAL problem with IPv4, is that there are many legacy products (for example the Xbox 360) which don't yet support IPv6. The suppliers of these products will say it's not worth updating as customers aren't demanding IPv6 as their ISP doesn't provide it.

It's not necessarilly a case of it not 'worth' updating to Ipv6, more often a case of not 'able' to do so! Most consumer level devices are made down to a price which necessitates providing the bare minimum to fulfil the task. For example, memory will be sized sufficient only to satisfy the particular requirement and so you can't necesarrily expect it to be able to accomodate 128 bit addresses when it's only be designed (and sized) to handle 32 bits. Dual stacking also requires additional storage which simply might not exist.

Quote

Because of these legacy products, ISP customers will continue to need to provide IPv4, irrespective of the IPv6 roll-out. Rather than implementing CG-NAT to extend the life of IPv4, surely the answer is for all the ISPs to rapidly provide IPv6.

Isn't that a contradiction in terms? We need both IPv4 and IPv6 to cater for the legacy products as you say and to provide room for future growth and improvement.

Quote

crucially customers would be hammering on legacy product manufacturers support helplines DEMANDING a firmware update for their Xbox or Roku video streamer, etc.

Customers can demand what they like but they won't necxessarilly get it. It might not be a) technically possibly, or b) commercially sensible. Upgrades of sold kit can undermine the purchase of new.

Quote

The final trick, would be for the clever ISPs to introduce a worthwhile monthly subscription DISCOUNT for broadband customers who use a connection that is IPv6 ONLY, reducing the pressure on the IPv4 pool.

Some are already doing that, or at least are charging a premium for IPv4 access using a public address (e.g. my brother's ISP in Luxembourg). An IPv6-only connection is little use to anyone though - you need both IPv4 and IPv6 if you want to access all parts of the Internet.

Quote

If this happened fast enough, then roll-out of CG-NAT might be avoided.  As it stands it all feels too little too late

It's never too late. The battle has been fought for years and whilst it might be nowhere near over yet giving up is not an option.

[#5]

My point was really that the ISPs need to jump first in providing IPv6, because without customers with IPv6 capable connections, there will never be a demand for change, either to existing or future products.
I can't figure out why the Xbox 360 console hasn't got IPv6.  There are so many people with games consoles complaining of NAT problems, or port forwarding issues, or more than one console on the same broadband connection, that if these consoles could use IPv6 then those widespread problems would be solved.  The website xbox.com is already IPv6 capable, and presumably Microsoft make a lot of money from subscriptions to Xbox LIVE - so Microsoft have an existing and ongoing revenue stream whether gamers upgrade hardware or not.

[#6]

My TV has an option to enter an IPV4 address via the remote control.
I wouldn't fancy typing IPV6 addresses in the same way.

[#7]

Your TV would probably just use Stateless Auto Configuration as would most other white goods.

[#8]

Quote from: purleigh

My point was really that the ISPs need to jump first in providing IPv6, because without customers with IPv6 capable connections, there will never be a demand for change, either to existing or future products.

It is indeed something of a chicken-and-egg situation. Content providers don't see a need to provide it over IPv6 because there are so few IPv6-ready consumers, and contect consumers don't see a need to ask for IPv6 connectivity because there's so little IPv6 (only) content. Hardward vendors are then stuck in the middle - why put investment into making their products IPv6 capable if they're not going to be used? The key to this circular dependency is for someone/something to break the circle... Running out of IPv4 addresses completely is sadly likely the only candidate.

Quote from: vilefoxdemonofdoom

My TV has an option to enter an IPV4 address via the remote control. I wouldn't fancy typing IPV6 addresses in the same way.

They're not actually that bad once you get used to them. Also, when statically configuring devices by hand it would be unusual to have to type the full address because thanks to address compression rules you'd usually pick an address with plenty of leading zero's in the host portion thus it wouldn't end up all that much longer than an IPv4 address.
e.g. instead of typing the full address 2001:0db8:0000:0000:0000:0000:0000:0001 you can simply type 2001:db8::1. In reality you wouldn't have the luxury of setting all those zeros given the smaller subnet size the ISP would give you but it hopefully illustrates the point.
The other beauty of IPv6 though is its support for a variety of auto-configuration methods including stateful DHCP like we have in IPv4 but also stateless auto-configuration too.
It would be highly likely that a TV would employ what of these techniques thus you shouldn't really be having to type in an IPv6 address anyway.

[#9]

Quote from: MJN

Whilst I'd be jostling to be the first in the queue for native IPv6 connectivity I can see why an ISP might keep quiet about their IPv6 rollout for commercial reasons. In a competitive market it usually pays to be one step ahead of your competitors, or to turn it around at least not be one step behind them. With an imminent depletion of IPv4 address space there could easily be ISP's, whose sole existence is of course to provide Internet connectivty (which requires address space to do so), who end up being unable to grow due to lack of address space. A company that doesn't grow dies. If an ISP is 'IPv6 ready' and can do so at the proverbial flick of a switch then it may see competitive advantage by waiting until the last minute to do so in the hope that others who are not ready get caught by the lion. I would not be surprised if this is what's happening - the collective 'herd' of ISPs may appear to be moving slowly on the subject but in reality they may well all be of the opinion that it's the slowest that will get caught - the rest just need to be going only slightly faster and they'll survive. That, or they're just clueless. And I honestly don't know where to put my money given that companies are usually controlled by the accountants and they might be planning only on a quarter by quarter basis as in the current climate the company might not be around to reap the rewards of any long term planning if it doesn't keep its head above the water.

I suspect the delay in the roll out is more down to routers, as you say 4 will still be needed which means effectively working out how to provide both 4 and 6 on a single router (irrespective of ongoing 4 issues) the tg582 requires  some major firmware rewrites to perform the required task without loosing functionality in other area's (like usb shareing).
then you need it to be customizable for each person, and then theres the cost of sending all your old customers a new router (which they would have to do for free) because there existing one is no longer useable on the network ......
all of that will be whats stalling there rollout (bean counters at there finest) waiting for old router owners to either order a replacement +12 mth contract or leave for another isp, sending out a new router at cost would kill there expansion ability for a year if those existing customers left ....

[#10]

Quote from: nanotm

I suspect the delay in the roll out is more down to routers, as you say 4 will still be needed which means effectively working out how to provide both 4 and 6 on a single router (irrespective of ongoing 4 issues) the tg582 requires  some major firmware rewrites to perform the required task without loosing functionality in other area's (like usb shareing).

AAISP has been providing a IPv6 service for over 10 years. Since the beginning of 2011 they have been allocating a /48 prefix to all new customers as standard. The router they supply? TG582
I've just spotted this on their website:

Quote

Non IPv4 connections Eventually we may have customers that do not need any IPv4 addresses at all. We already have support for NAT64 to allow IPv6 only networks to access Legacy IPv4 machines on the internet, and we will prefer this as a solution than doing carrier IPv4 NAT. This is something people can try now, and by the time it is needed, some years in the future, the technology should be mature. However, we expect more and more of the services on the internet to be IPv6 by then and legacy IPv4 access to be an niche requirement.

I hadn't heard of NAT64 before.

[#11]

indeed,
which would support the idea that non tg582n's would need to be phased out before complete 6 rollout, nat 64 is a cute way of saying auto tunnel 4>6 and have it work at a device level as something that can connect, its not new and its definably better than the overly complex cg-nat idea, unfortunately its not much use in terms of mitigating the problems of running out of ipv4 addresses (you still need one at the gateway)
if an isp phased in pure ipv6 and gave out routers configured for dual stack with auto wrapping of legacy clients then they wouldn't even need a nat64 gateway, the difference is of course the gateway is a huge money pit, the dual stack routers would only be required for 10>15 years but the cost of providing them would prove almost as high as the gateway if more than 80% of customers required them, currently only a few sub £200 routers support this option which makes it economic suicide to chose it over the relatively cheaper rollout cost of a gateway nat64 solution, it all depends on what space they have in there data centre as to what they will end up doing eventually 

[#12]

You will always need some kind of ipv4 capability.
I understand that a few computers are still connecting via dial-up?  
BTW Would ipv6 still work over ADSL?

[#13]

Quote from: vilefoxdemonofdoom

BTW Would ipv6 still work over ADSL?

Quote from: jelv

AAISP has been providing a IPv6 service for over 10 years.

Some (most?) of the Plusnet IPv6 trialists were on ADSL.

[#14]

it did when I used to be a BT customer in the trial area, the isp servers just assigned a 6 address and the router did the rest, apparently it didn't work for most people though but then they were using home hubs and apparently they couldn't do the do...mind that was several years ago

[#15]

If you meant dialup rather than ADSL then the answer is still yes - it uses PPP in much the same way as ADSL does.
In the OSI 7-layer model ADSL and dialup are operating at layers 1 and 2 whilst IP (v4 and v6) sits at layer 3. This level of abstraction and decoupling means the underlying carrier shouldn't matter. In theory at least; in practice you need to consider aspects such as how address assignments are made between the two hence the layer 2 protocol does usually need to be 'layer 3 aware' and thus would need to be explicitly configured/operated to support IPv6 over it.