The inverse situation (IPv6-only client but IPv4-only server) is not really an issue, since for that situation NAT64 works, since you can embed IPv4 addresses into IPv6.
The only way C (ipv4-only) could communicate with S (ipv6-only) is by either allocating a dedicated ipv4-address to S (doesn't have to be directly connected to S - it can be sent to some translation box that does SIIT) or by upgrading C to support ipv6 and tunneling it (6to4, 6rd, teredo, etc.).
If S is IPv6-only, then either the address is from the IPv4 range, or it's not. If it is, then R can translate between them. If not, then they can't connect. I don't see him suggesting otherwise anywhere.
Even ignoring translation possibility, his point here isn't to claim its magically possible for a server outside the IPv4 address range to talk to an IPv4 client. His point is that this can move everyone to IPv6 for communication, without ISPs also having to deal with leasing addresses outside the IPv4 range, which he believes would help adoption.
But maybe I should reread the article a bit more generously and less literally, maybe he is actually advocating for the IPv6 we have today? Because there are transition technologies like NAT64, SIIT, 6to4, 6rd, etc. that sort of allow some of the things he suggested? "If the IPv6 configuration isn't automatic, it won't happen." sounds like he is actually advocating SLAAC to me?
> What else could he have meant with "In other words: The current IPv6 specifications don't allow public IPv6 addresses to send packets to public IPv4 addresses."
He could've meant "making IPv6 work means much more than upgrading software. Every administrator of a server on a public IPv4 address has to go to extra effort to acquire and enable a public IPv6 address."
You cannot read this and miss the fact that his frustration is with the need for administrators to acquire separate IPv6 addresses.
> maybe I should reread the article a bit more generously and less literally
No need to do either. Just read all of it, instead of 1 or 2 sentences that might sound like amateurish mistakes when you remove several pages of context and then extrapolate without them.
Adding the address to the server is the easiest part of the whole process. Upgrading the software is probably the biggest hurdle with people just hardcoding things to 32-bit left and right (and some badly designed APIs which do either IPv4 or IPv6 but not both in a transparent fashion). And next is actually setting up ipv6 connectivity.
All those other things you would still have to take care of even if you magically could keep using the IPv4 address in IPv6 (however that would work). Never mind that servers don't statically sit around forever, but one has to set up new servers all the time.
So why not go for the inverse approach for new servers? Only give the servers IPv6 addresses and for those servers that still need to be reached via the public IPv4 internet (most internal ones probably have no need of this) can be accessed via SIIT, which performs stateless mapping between IPv4 and IPv6 (for example IPv4 1.2.3.4 gets translated to IPv6 2001:db8::1.2.3.4 and vice versa).
You don't need to agree with me on what he's saying. The quote "...extra effort to acquire and enable a public IPv6 address" is a directly copied from his own page. It doesn't make sense to read that and claim that's not what he's saying, or to claim I'm "interpreting" it to mean that. There isn't really any other option than to go with that interpretation; that is what he's saying - it's his quote.
> Adding the address to the server is the easiest part of the whole process.
The issue isn't "adding the address to the server". The issue is obtaining said address to be able to add it to the server in the first place. This requires both you and your ISP to set up and manage/maintain an entirely independent parallel network with an entirely separate configuration. It's an administrative hurdle, not merely a technical one. I can't explain it better than this user did, so just read his comment: https://news.ycombinator.com/item?id=37555424
In any case, it's fine if you disagree that this is actually a significant hurdle; I'm not trying to argue that point. My goal here was to portray what djb is saying accurately, not to agree or disagree with it. So if you disagree with it, that's mission accomplished (for me anyway).
Funny, because the reality is that the software has for the most part been written and the hard part is acquiring v6 addresses. The problem is it isn't just a matter of acquiring a v6 address, it's also supporting the software that the v6 address runs on, and having it work the exact same way as v4.
The answer is that I'm actually talking about a huge number of IPv4 sites. There's nothing special about my site. When the same situation is repeated at N sites, the code is written only once, while the trivial requests and the trivial configuration changes are repeated N times.
.
.
.
Well, I'm looking at a much larger group of users, and most of them aren't putting even five seconds into IPv6. They have better things to do. If the IPv6 configuration isn't automatic, it won't happen.
> So why not go for the inverse approach for new servers? Only give the servers IPv6 addresses and for those servers that still need to be reached via the public IPv4 internet (most internal ones probably have no need of this) can be accessed via SIIT, which performs stateless mapping between IPv4 and IPv6 (for example IPv4 1.2.3.4 gets translated to IPv6 2001:db8::1.2.3.4 and vice versa).I'm unfamiliar with SIIT, but I bet it doesn't get us closer to the "magic moment" or we would be flocking to IPv6 addresses:
The magic moment for IPv6 will be the moment when people can start relying on public IPv6 addresses as replacements for public IPv4 addresses. That's the moment when the Internet will no longer be threatened by the IPv4 address crunch.How are R and S negotiating when R cannot even name S on the network? Its stack only allows for 32-bit addresses and S can't have a 32-bit address.
Let me summarize my understanding of what he's saying, because I don't quite see why/how you disagree. I think you (or I) might be misunderstanding his claim.
Imagine this topology: C (client, IPv4-only) <=> R (intermediate router) <=> S (server)
My understanding of djb is he's saying that IPv6 could have been designed such that S could still serve C via only simple software updates -- this means, crucially, without the need for S to separately obtain a public IPv6 address through R, because its IPv4 address would be automatically valid for IPv6.
How can this work? Well, there are two scenarios:
1. If R is IPv4-only, then S could figure that out during some startup/negotiation process, and send only IPv4 packets to R. R only lets IPv4 clients connect to S anyway, so any response (even from IPv6 applications) on S must be going back to an IPv4 address. So the kernel can transparently translate those IPv6 addresses into IPv4 before passing them along to R, and vice-versa.
2. If R supports IPv6, then R can do the same thing S would've done in the previous^ scenario. (In fact, I think S could become IPv6-communication-only in that case, reserving IPv4 for just address leasing? I'm not sure, but in any case, I don't think that matters here.)
Notice that all of this is almost completely stateless. (I think the only state S needs to track here is 1 bit, indicating whether R supports IPv6 or not.) So, S and R can be independently and (importantly, rather trivially) upgraded to support the IPv6 protocol, without losing the ability to talk to any clients within the IPv4 address range.
This is easy and requires no explicit leasing of IPv6 addresses. That step can be implemented and have support for it added later, whenever S is ready to serve clients beyond the IPv4 address space.
Does this make sense? If so, then it seems to show how IPv4-only clients could talk to IPv6 servers without modification. If not, then I'd love to see where I'm mistaken (I very well might be).