Hacker News Reader

Comment by tyingq | Hacker News Reader

I wonder if they will implement some kind of clawback for orgs that are sitting on large chunks of sparsely used blocks. Old tech companies and colleges come to mind.

https://myip.ms/info/ip_ranges_blocks/Major_Biggest_IP_Addre...

View on HN

I doubt that, but I think we're going to see more and more companies to sell off the unused blocks. Last year MIT sold off 1/2 of 18.0.0.0/8 to Amazon.

It looks like the going rate per IP is between $20-$24 currently:

https://auctions.ipv4.global/prior-salesw

That's a good point...further scarcity will drive the prices to a point where selling is very attractive.

ipv4 is the new bitcoin?

Don't know how credible it is, but these folks were predicting the current ~$20 price goes to ~$50 over a year's time:

Chart: https://www.retevia.net/wp-content/uploads/2019/02/IPv4-Pric...

Article: https://www.retevia.net/address-pricing-2019-and-beyond/

Yeah, I think there will be a small bubble as some try and buy all they can while we continue to transition to ipv6

Has a date been set for when IPv4 will stop being routed?

No, because the Internet is a voluntary system.

Very few companies would need to decide to not interconnect with IPv4 networks before IPv4 was effectively dead. Typically, though, there's a process that they would go through before making that move. That's what I was asking about. Has anybody floated a date for IPv4 death?

Those very few influential people who make the decisions are incredibly conservative; they won't even kick criminals off the Internet without overwhelming evidence. I don't think they're even interested in discussing IPv4 sunset at this point.

Rather than a date I think it would make sense to look at the traffic. When IPv4 hits, say, 1% it might make sense to turn it off. I would estimate that's decades away.

Google is one of those influential companies and I could totally see them declaring that their crawler is going to stop visiting IPv4 after some date.

They would do that only once it costs them more money than they could make with the retrieved data. That's not going to happen anytime soon.

You might as well wonder whether google is actually going to stop crawling IPv6 addresses (are there any significant IPv6-only servers?)

That's true. However, even if they just said IPv6 destinations will be used in the ranking algorithm as a positive indicator of a site that's being actively maintained, that alone would be enough to get some people moving.

It's a big gamble on when to sell though. Eventually IPv6 adoption will hit a critical mass threshold and interest in paying for IPv4 IPs will dry up.

GE did the same thing. It sold 3.0.0.0/8 to amazon about a year ago.

Who is 'they', and under what authority? If you're thinking about RIPE or ARIN, legacy block allocations predate those organisations even existing. They have no rights over that space.

At the same time, IANA has delegated 'future' (read: sometime in the past, but not until after those legacy allocations) responsibility to the RIRs (ARIN, RIPE, APNIC, LACNIC, AFRINIC).

Finally, how do you determine whether or not space is used? Ping? BGP announcements to the Internet of the space? Neither of those concludes that the space isn't used, and RIR policies permit for registration of globally unique addresses without any obligation to announce to the Internet.

Not sure on who "they" is, to be honest. But if/when ipv4s become very difficult to obtain, lots of potential "theys" will likely emerge. Powerful companies and governments won't just shrug it off.

Similar to eminent domain maybe.

At some point the cost of doing all of what you mention is more than the cost of NAT64/46 infrastructure to enable the problem to actually be fixed for more than a couple of years at a time.

When the options are: let them profit awkwardly from a windfall, claw it back against their will, or make their IPv4 assets obsolete, focusing on the latter seems most productive.

The UK’s largest mobile provider now gives me an IPv6 only connection and I didn’t even notice.

Apple have made it a hard requirement for apps to operate correctly on IPv6-only connections for a few years now. If your app doesn't work unless it's on IPv4, Apple will reject it.

Pretty sure that doesn't mean all distant API endpoints need ipv6 addresses.

Just means no hardcoded ipv4 or incompatible API calls. So that ipv6 to ipv4 gateways work.

So, the app is ready, but apple isn't dictating much about your backend.

Apple's mandate happened because a couple US Telecoms asked them to. The US has plenty of cell towers that are IPv6-only, and at least a couple US carriers are now IPv6-only or IPv6-primarily, with more planning to switch. All the carriers have big NAT64 gateways of course to keep the internet from entirely breaking, but that's still an avoidable bottleneck from a backend perspective. Apple themselves may not be dictating what the backend is doing, but they are doing a lot of Happy Eyeballs [1] tests on your backend and IPv6 without NAT64 proxies in the way is likely the fastest experience for a lot of real mobile users today.

[1] https://en.wikipedia.org/wiki/Happy_Eyeballs

That can't be totally true, Twitter doesn't work on v6.

Apps can't request IPv4 addresses directly (ie, no hardcoded IPv4 addresses allowed), but DNS may still return them for indirect use (via the Happy Eyeballs [1] approach that heavily favors IPv6 because an increasing number of mobile carriers are IPv6-only or IPv6-primarily and IPv4 addresses must get sent to NAT64 gateways).

[1] https://en.wikipedia.org/wiki/Happy_Eyeballs

Not only can it be true I've done it at a conference, which is to say I've implemented an IPv6 only network for conference delegates and had zero connectivity complaints.

The ingredient you are missing is while the clients only talk IPv6, the gateway they go through (must) have both IPv6 and IPv4 addresses. There is a well defined mapping from IPv4 to IPv6 addresses, so if the IPv6 clients need to communicate with a IPv4 only host they use it's mapped IPv6 address. The gateway then NAT's all those mapped IPv6 addresses to it's IPv4 address.

Which only leave the problem of does an IPv6 client know to use an mapped IPv4 address, given it is supposed to know nothing about the IPv4 world. The answer is you make it use a DNS server that lies. I'll leave it to you to figure out exactly how it lies, but it isn't hard to figure out given it has clients that only understand AAAA records, and some hosts (such as twitter) that only resolve to A records.

> The gateway then NAT's all those mapped IPv6 addresses to it's IPv4 address.

Right, then its not a pure v6 network, its got v4 somewhere patched in. When I'm on my pure v6 network at home, Twitter does not work.

And yet one of the largest UK ISP's by customer count (Virgin Media) still haven't managed to roll out DS-LITE ... 6 years after announcing it.

The major browser currently hides "www" and " https:// " by default.

Now it has to hide port numbers and IPV4 space will suddenly be multiplied by less than 2^16.

When I was at Cisco, I'm pretty sure my desk phone had a public IP.

In what, 1998? They have been on RFC1918 for like 20 years.

Source: Worked at Cisco.

As far as I understand from previous discussions, those blocks (1) can't be split among a bunch of providers, since it would bloat routing tables and (2) wouldn't extend the timeline very much, so it's not worthwhile engaging in that fight.

There's apparently a wait list for /24's now. Desperation might inspire a new set of discussions.

I believe BGP tables are close to crossing a threshold. (768k entries) that breaks a bunch of routers, so they need to address that anyway.

We're already past 768k and are now around 800k[0].

[0] https://bgp.potaroo.net/as2.0/bgp-active.html

There's another problematic threshold at 1024k.

HP have two contiguous class A 15 & 16 the latter was inherited when they took over DEC. That cant be justified.

Reply on news.ycombinator.com