IPv4 is almost gone (in Europe anyway) | Hacker News Reader

IPv4 is almost gone (in Europe anyway)

ripe.net

78 pointsmstevens52 comments

This seems to have been rumbling on for some time now - at least for the past five years or so.

My OS supports IPv6, my home router supports IPv6, but at no point have any of the home ISPs I've used (three in the past five years - BT Internet, Be and Origin Broadband) made any mention of any sort of IPv6 support.

I'm not au fait with the workings of such huge operations, so is there any reason for ISP adoption of IPv6 to be so sluggish? It's not like it's been an overnight thing. I'm aware it's unlikely to be a simple case of flipping a switch, or installing a new software package - but we do appear to be approaching an IP crunch.

Are there any consumer ISPs that do offer customers a block of IPv6 addresses for use, rather than (or as well as) a single IPv4 address?

I'm in the opposite situation.

My ISP and wifi router support IPv6; but my modem does not. And without this modem failing, I probably won't shell out $150 for a new one.

According to [1] there are a few - no big names, though. As [2] says, running IPv6 is fairly boring; you get access to the same web as everyone else.

[1] http://www.sixxs.net/faq/connectivity/?faq=native&countr... [2] http://bens.me.uk/2011/adventures-in-ipv6

I'm running on IPv6 on Comcast in the US, and yes while I get access to the same web, I have noticed that I was having issues with IPv4 that I don't have on IPv6. Such as limits per IP/account tuple logged into certain servers.

So my iPad, iPhone, and MacBook Pro being logged into the same Google account over IMAP would sometimes cause issues because of too many connections for that specific account, now with IPv6 that isn't an issue because all of my devices now have their own unique IPv6 address. Perfect.

[2]'s comment "you get access to the same web as everyone else" really only applies to consumers of "the web".

And - more importantly - soon no longer applies even to that. IPv4 is essentially gone, and when it is, not having IPv6 will mean your web is smaller.

IPv6 gives you much more freedom in configuring your network as both a service provider and consumer. This isn't boring :)

Eircom (an Irish ISP) are trialing IPv6 currently, and the routers they have been shipping for the last year or so support it.

Other than that, our co-location provider Blacknight offer IPv6 to their customers.. but that's not exactly residential!

BT / Esat were trialling IPv6 as early as 2005 and giving native allocations around that time also

> This seems to have been rumbling on for some time now - at least for the past five years or so. ... It's not like it's been an overnight thing.

To say the least! This particular rumble ("we're out!") may be new-ish in the last few years, but the general rumble ("we'll run out!") has been going on a lot longer, and the transition is pretty much the opposite of an overnight thing. I learned about how IPv6 was imminently needed and right around the corner when I took a networks class in college. In 1996. Tanenbaum's book (from four editions ago!) includes ten pages on IPv6 that could appear in a book published today nearly unchanged.

My ISP (Andrews&Arnold, UK) offer IPv6, so I have a /48 and (since I bought a new modem recently) all my machines are now natively IPv6 with no tunnelling involved. (A&A also support tunnelling if your modem/router is IPv4 only.)

Excellent. I currently use a local ISP, because it's the only way I can get a decent speed where I live (there's a local FTTC network called Digital Region).

But I'm likely to be moving out of their coverage area at some point, so I'll add this to my list to consider when I do. Thanks!

> Are there any consumer ISPs that do offer customers a block of IPv6 addresses for use, rather than (or as well as) a single IPv4 address?

Comcast, at least in some parts of the country (e.g. SF's East Bay). You don't even have to ask them. One day I noticed my router was being advertised an IPv6 route. I switched on DHCPv6 and got an IPv6 address, as well as a /64 prefix for the rest of my network.

Unfortunately /64 is all you can get for now, which isn't really sufficient, but they've promised to delegate shorter prefixes in the future.

"Unfortunately /64 is all you can get for now, which isn't really sufficient"

Not sufficient for what? (Honest question.)

yeah, a /64 is more than 2^32 times more than all of current internet... i don't think the GP really meant 'sufficient', or i'm missing something.

a /64 gives you a single subnet with effectively infinite size. A shorter prefix would allow you to have multiple subnets for isolating things like gaming devices, guest networks, VPNs, etc. For IPv4 the ability to do this was basically a pipe dream, for IPv6 the addresses are there, we just need to coordinate how ISPs hand them out and how consumer routers manage them.

In short, one /64 is sufficient for duplicating an IPv4 like situation. A /60 or /56 gives us more room to innovate in new types of home networks.

Sufficient to have more than one subnet. In IPv6 you can't easily have a subnet that's smaller than /64, so even though /64 = 2^64 addresses, it's still just a single subnet. In IPv6, we need to start thinking in terms of number of subnets rather than number of IP addresses.

Why is having one subnet insufficient? I use multiple subnets (I have a guest network), and I don't want to have to use NAT with IPv6 when my ISP has effectively infinite address space. Of course, I'm not an average user, but remember that 10 years ago only power users had home routers, and for everyone else 1 IP address was enough. Now everyone has a router. There are already some consumer access points that tout a "guest network" as a selling point, and there may be other consumer uses for multiple subnets in the future.

a /65 is a subnet, and you get TWO of them with a /64. And so on. So I don't really understand what you're talking about here.

In theory, you're right. But in practice, IPv6 really wants subnets to be /64. For example, stateless address autoconfiguration requires it.

It truly is saddening how we're treating v6

I agree. It seems like we're setting up the IPv6 address space to be wasted.

Actually, there's a good reason why IPv6 subnets are so large - it's because with stateless auto configuration, the lower bits (the "Interface ID") of the address are derived from the MAC address, which is 48 bits.[1] (Stateless auto configuration is brilliant because, despite being stateless (unlike DHCP), no two devices will ever be assigned the same IP address on a subnet.) 48 < 64, so technically the subnet size could be /80 instead of /64, but does 16 bits really matter in the grand scheme of things, especially when there are 2^64 /64's available?

Another way of looking at this is that IPv6 could have been just IPv4 with 64 bit addresses, and we would have been fine. However, the designers decided to add stateless auto configuration, and added another 64 bits to make it work.

[1] In privacy mode, the Interface ID is generated randomly instead of from the MAC address (to prevent tracking by MAC address). In this case, you still need lots of bits to make the probability of a collision insignificant.

NAT will never disappear.

You're probably right, and one easily preventible reason will be ISPs giving out too-small or dynamic prefix assignments. There may be legitimate reasons for IPv6 NAT, but this isn't one of them.

I'm not sure about other places, but I wrote some software for an ISP in Poland. It seems damn hard to coordinate whole country worth of people in normal operation, leave alone make a transition to new technology. Remember that many people won't be happy with transient trouble related to the switch.

I'm afraid of the possibility that instead of adopting IPv6, internet providers will just switch to ISP-wide NAT, making the internet even more asymmetric than it already is.

I sincerely hope not! ISP wide NAT is bound to cause countless issues for internet services and businesses.

The first and mose obvious examples I can think of:

* Internet services (eg websites, netfix etc) can no longer blacklist IP addresses without blocking entire ISPs.

* Businesses can no longer offer "direct to the office" VPNs for remote workers.. Actually - even site-2-site VPN's will break if both sides are behind an ISP wide NAT.

Look on the bright side: users will have more privacy.

http://panopticlick.eff.org/

Works over NATs and proxies. And obviously the techniques from there are already used by those who want to sell your info for profit. So, please, let's put this meme to rest :-)

Blacklisting can and will happen based on other things, just that it is more costly and less performant. So the consumers will pay for that in hidden costs - less of the "useful" services delivered, etc.

The VPN part is sort-of correct - it will depend on the type of the NAT. With most of the NATs, establishing the direct connection over a pair of them is technically possible - take a look at STUN, TURN and ICE (IETF standards).

The first is a feature. The second is not true: a VPN can tunnel through TCP in the worst case, and most NATs allow VPN over UDP using some tricks.

To be clear: transitioning to IPv6 is a two step process:

1) you need to enable IPv6 everywhere 2) when that is working, you can sunset IPv4

So, every transition strategy involves both the life support for IPv4 and the delivery of IPv6.

Sometimes the ISP has no choice but to use it. And iff the ISP also rolls out IPv6 in parallel, this is fine, the users can access the dualstack websites over IPv6, while still using IPv4 for the rest of the content. And this is the key.

A lot of folks here are building products that have their front door on the Web.

And this is where all of you, as a "content provider", can make a difference for those users who already have IPv6 by providing the service both on IPv4 and IPv6.

Agreed - ISP wide NAT will be necessary, but if they choose that instead of ISP wide NAT+IPv6, then we're in trouble!

EDIT: Also - Everything I have a hand in is accessible via both IPv4 and IPv6..

In my conversations, the vast majority of the folks who are going to have to deploy CGN, also have already running programs to deploy IPv6. It's just that they don't necessarily advertise them, and also that these things take years - the access is by far the hardest part of the internet to transition.

At my $employer, a couple of bright interns this summer made this: http://6lab.cisco.com/stats/ - which aggregates the stats from various places and allows to get a more rounded view. Besides for the access part (EDIT: which, as I said, just takes some time), the situation is not bad at all!

EDIT: And, kudos to dualstacking your gear! :-)

From what I understand (which admittedly is not all that much) implementing so called carrier-grade NAT would be almost as expensive as transitioning to IPv6 would be. And even private ranges are finite, which afaik was one cause for Comcasts applaudably early transition to IPv6

For those of you curious about IPv6, and wanting to experiment with it at home or in the office, have a look at SixXS[1] and Hurricane Electric IPv6[2].

They both offer free IPv6 "tunnels" which can be used to provide IPv6 to your home/office.

We use SixXS in the office, and the tunnel has been alive without issue for 83 weeks and counting..

(P.S. SixXS is probably the easier choice if you have a dynamic IP, or are not setting the tunnel up from your router directly..)

[1]: http://www.sixxs.net/ [2]: http://ipv6.he.net/

worth noting that HE has tunnel endpoints all over the world. i'm using one right now in europe and it works really well.

I know these folks are well-meaning but I find these constant "warnings" to be insulting. What people really need to be aware of is that the folks making decisions of how the internet should run, "the experts", are not always as smart as they think they are. They make mistakes. And it's hard to get a bunch of know-it-all's to agree.

It's also hard to get the entire `net to switch their behaviour, and adopt something that is incompatible with IPv4, without telling them what benefit they will gain by doing so. Can you blame them? But the "experts" and their zombie followers sure are trying.

IPv4 works. NAT works. NAT can be traversed. So what is the problem exactly?

As a home user, I can set up my own NAT'ing scheme with private IPv4 space far easier than I can learn to deal with the added complexity of IPv6.

OK, now I will get skewered by IPv6 fanatics. How many of them are getting paid to do IPv6 consulting work?

IPv4: It Just Works. "Experts": They make mistakes, just like everyone else.

Now, let's hear from the "experts".

What happens when your ISP has to NAT you?

And in Asia: APNIC initiated last block measures in April last year.

http://www.apnic.net/publications/news/2011/final-8

"almost"?

I'm still stuck on an IPv4 connection at home, in Britain. My phone still can't get IPv6 on mobile internet. My VPS host still doesn't support IPv6 on my server. And PHP still fails to properly parse some types of IPv6 address.

Things have got better, sure, but don't say it's almost gone. Far from it.

The notice is about IPv4 allocation, not about IPv6 market penetration. RIPE NCC has just begun allocating IPv4 addresses from the last available /8 block.

Ah, fair enough.

> PHP still fails

It still fails to reject February 31, I'm completely not surprised to hear that it fails on IPv6.

I don't know which specific function you're talking about, but there may be a good reason for it.

Both strtotime() - PHP's "take a string and do your damnedest to make a date out of it" function - and DateTime.parseFromFormat() - the more restrictive TZ-aware class - accept it and give you March 2/3. There's a separate validate_date() or something like that that actually rejects Feb 31.

I'm not sure that's entirely wrong, just a different attitude than you might expect. Also, for that matter, I think MySQL accepts, will store, and will retrieve Feb 31 IIRC. Why? It's liberal in the correctness of input.

Hey 2004 called, they want their headline back.

When you consider the ability for inter-RIR transfer of networks, not to mention the market being developed (already released?) by ARIN (do other RIRs have markets?), and you've got a great recipe for a run on IP addresses.

IPv6 is just like the transition to Python 3: it'll happen... eventually. But the only people with an incentive to put the hard work in are people who are genuinely enthusiastic about the transition. Everyone else is sticking it in the too-hard basket.

This is a fairly outdated meme by now. Just that the enthusiasts are loud about it. Everyone else does not rush tell the world about their plans - it takes on average about 3 years to get everything in place, so this is a pretty good head-start against the competition :-)

I didn't realise I stumbled upon a meme!

Yep and nothing will be done by ISPs until they are all completely gone (and maybe not even then, scarcity is good for business).

This isn't true.

Comcast, Verizon, and T-Mobile are a few of the big eyeball networks who are making large infrastructure investments in IPv6. ISPs are far ahead of most content and hosting providers in this area, as they bear the pain of IPv4 scarcity when providing addresses to their subscribers.

Transit providers are on the whole prepared for v6 as well.

As an ISP, you don't want to wake up one day and see analysis that your network numbering is at odds with your growth plans.

Not only are they making large infrastructure investments, a many of their customers (including myself) can get IPv6 service from all three of those companies today and have been able to for some time.

Reply on news.ycombinator.com