Comment by runjake
5 days ago
> still hasn't taken over the world
Maybe not in the strict sense, but it kind of has.
In the enterprises I've worked in the past decade with IPv6 running, at least 75% of the Internet traffic is IPv6. In my discussions with other engineers managing large networks, they seem to be seeing more or less that same figure.
The problem is that virtually nobody knows IPv6. I regularly bring up IPv6 in engineers' circles and I'm often the only one who knows much about it. And so, I have doubts about it's long-term future, except for edge cases. I figure some clever scheme utilizing IPv4 and probably NAT will come around at some point.
IPv4s are about to be bought, held, portfoilo'ed, speculated, and rented/mortgaged/sold like real estate. Companies like IPXO are already doing it. The costs of public IPv4's are going to go up for no technical reason because a new distinct ownership layer is springing up between you and the ISP. You're going to start renting them or paying a holder for the right to use them (on top of your ISP to transport it) at some point. And you can continue to do that, or get IPv6's for free.
Just to be pedantic, it's "illegal" to hoard IPv4 or to buy it for any purpose other than using it directly. But yeah, in the real world it may become more financialized than it already is. OTOH if prices keep dropping maybe they won't bother.
Ford Motor Company has both a /8 and a /9. They own over 16 million ip addresses.
Relatedly, I've been seeing some people buying up old domains and squatting on them with AI generated content. Not even ads, but content that seems like something that might actually show up in a rare Google search query. Not really sure what the play is or why this is better than advertising the domain for sale (do registrars punish overt squatting these days?).
IPv4s have been bought and sold for years
https://auctions.ipv4.global/prior-sales
Prices have been going down in nonimal terms for years, let alone real terms. In terms of investment they're a terrible asset.
IPv6 and CGNAT growth has finally started to suppress IPv4 prices. There was a huge pump when hyperscalers decided they needed more. But IPv6 keeps growing and is the majority of traffic in many networks. If you own significantly more IPv4 addresses today than you need, I would dump them on the market yesterday. Spend some of the profits to move to IPv6 if still needed.
nice. I wish I could buy an address instead of renting from aws...
It seems like the addresses cost about $20 each, and can be rented out for ~$5/year.
That doesn't seem terrible.
I'm a networking noob, but would it be possible to extend DNS/HTTPS so as to allow a URL to point to a port other than 443? Doing so would allow each IP address to serve multiple websites/computers making the pool of addresses at least thousands of times larger.
As others have mentioned, there's SNI and host headers to have multiple sites on port 443, but there is also the SVCB/HTTPS aliases (https://www.rfc-editor.org/rfc/rfc9460) which will allow having the plain domain alias to other hosts including ones with embedded port numbers. Non-browser support is pretty lacking though.
That’s sort of what HTTP is already doing though no?
Multiple websites can have the exact same DNS record and live on the same physical server / IP address, but the HTTP(S) request must specify what host name it is actually requesting, so the server knows how to serve it.
It is already possible using the Host header and TLS SNI. But traffic still flows through port 443.
How does one get an IPv6 allocation for free? Or, do you mean the ULA space? Because the latter doesn't really count.
You just ask your RIR. For example: https://www.arin.net/resources/guide/ipv6/first_request/
15 replies →
We own our own IPv4 and IPv6 ranges, which is nice. There already is a holder for the US: ARIN.net and I hear it's a pretty spendy annual fee for most orgs (we're legacy. we've had ours for decades)
Now all we need is for someone to make a crypto currency so you can fractionally own IPv4 addresses.
Presumably this would be port-based fractional and 443/tcp would cost a premium.
2 replies →
> Maybe not in the strict sense, but it kind of has.
I challenge you to find:
1. A hotel in the US that provides IPv6. I have NEVER been in one, and I once stayed in a hotel (in Mountain View, CA) that was giving out public IPv4 addresses.
2. An easier task: a SIP provider that has IPv6 (in the US). You know, for the VoIP that is supposed to be a poster child of end-to-end connectivity.
> In the enterprises I've worked in the past decade with IPv6 running
What about those without IPv6 running?
Anyway, in the enterprises I've worked in the past decade - of course, another anecdote - not once has anyone ever specified an IPv6 address of anything. Inside the organization or outside of it.
why would an enterprise turn to IPv6?
everything fit's nicely in the 10.0.0.0/8 range
in my many decades of enterprise infrastructure, no-one has ever mentioned IP6 either.
why would they, whats the business case?
The problem with private address ranges is that everyone thinks they're available. In a large enough enterprise you're bound to have conflicts. They usually pop up at the most inconvenient time and suddenly you're cosplaying ARIN in your IT department.
> everything fit's nicely in the 10.0.0.0/8 range
Except during a merger/acquisition and both companies have 10.0.0.0/24 in their OSPF or IS-IS topology.
> everything fit's nicely in the 10.0.0.0/8 range
Except for when it doesn't.
If you just use that space as a flat range, it is almost certainly more than enough. But if you split it up in multiple levels of subnets, you can run into difficulties balancing having enough subnets and having enough space in each subnet.
3 replies →
We burned thru pretty much all of our public /8, RFC1918, and have begun digging into RFC6589 (a /10 I didn’t even know existed prior to job). Still shocks me. Hardly an expert in the space, but I think the issue comes from subnetting to distribute ranges to teams that need a consistent IP address space for some project or another. Lots of inefficiency & hoarding over time. We’ve had legitimate outages and impending platform death staved off by last minute horse-trading & spooky technical work due to such things. IPV6 has always been a distant aspiration.
Grow large enough and you hit the limit pretty fast. NAT complicates things.
2 replies →
you haven't had to set up intercompany vpns I see
5 replies →
Unless you get to big. Or you merge with another company and have to combine your internal networks and oops, all the subnets are overlapping. Or you need to serve mobile clients who get better connectivity over v6.
if both you and companies you have site to site vpn with have IPv6 there is no IP conflict or NAT to worry about.... and that's about end of the advantages
one poorly made decision and oops you're out of 10/8 addresses
if you've never run in to this, then sorry, you've not been in an enterprise, you're in a mom 'n pop shop cosplaying as enterprise.
> not once has anyone ever specified an IPv6 address of anything. Inside the organization or outside of it.
If you deploy IPv6 correctly, you shouldn't have to disclose IPv6 addresses to users inside or out -- DNS keeps the address literals abstract, hidden from users.
I am on my company's VPN right now and I get a 0/10 at test-ipv6.com
>Maybe not in the strict sense, but it kind of has.
>In the enterprises I've worked in the past decade with IPv6 running, at least 75% of the Internet traffic is IPv6.
Nobody cares about those. What matters is if my device has an IPv6 address assigned.
Ok then: most people in the US do. The rest of the world is looking increasingly ipv6 too: https://www.google.com/intl/en/ipv6/statistics.html#tab=per-... India is 71% IPv6 (probably thanks to Jio), China has it in its 5 year plan, Europe is doing well, etc
Wasn’t it mandated for 4G? Or at least 5G?
1 reply →
> at least 75% of the Internet traffic is IPv6.
> Nobody cares about [that]. What matters is if my device has an IPv6 address assigned.
This seems to be the weird dichotomy in these comments. Some people are arguing from the position that is absolutely everywhere and is doing great.
Others are saying since their machine doesn’t show it it’s dead and no one cares.
Is there a term for this? A successful failure? A failed success?
Kind of odd.
It is why the Google IPv6 stats fluctuate between weekends/holidays and weekdays. IPv6 is much more prevalent on home and mobile networks so increase on non-work dyas. Companies have IPv4 networks that they don't want to upgrade. We have dichotomy where 50% of clients have IPv6, but most of the small sites do not.
The other thing I have seen is that engineers make things complicated. Normal person has IPv6 enabled by default or enables it in router, and it just works and they never notice. Engineers want to configure things manually, but IPv6 is hard if fight against the dynamic defaults.
Maybe the False Consensus Effect?
https://en.wikipedia.org/wiki/False_consensus_effect
Anecdotal stalemate.
I use this argument, because HN also tries to do the reverse when someone suggests a protocol/addition/replacement to either TCP or HTTP. Then suddenly it's important what shitty company networks do. It's still not.
75% or 99% does not matter. Until you can't forget about IPv4, IPv6 us useless.