Talk:IPv6

Wikipedia IPv6 enabled

(There was only a pre-existing heading here and no text or comments. My note here was in response to the heading.) Note that while the web servers for www.wikipedia.org have IPv6 addresses, it seems that the authoritative DNS servers do NOT have IPv6 addresses. This means someone on an IPv6-only network would not be able to look up the addresses to connect to Wikipedia. I noticed this when www.wikipedia.org failed the Internet.nl website test for IPv6. I do not know who to report this to or how to make that report. - Dyork (talk) 01:29, 8 March 2020 (UTC)

Try WP:VPT. Johnuniq (talk) 02:05, 8 March 2020 (UTC)

 Done(in terms of the issue being reported) Thank you for that suggestion. I have noted the issue there. - Dyork (talk) 02:36, 8 March 2020 (UTC)

• @Dyork and Johnuniq: phabricator:T81605 Long story short is, the Foundation wishes to first get IPv4 DNS on anycast, then get IPv6 support ready. This inherently gets around the problem of DNS geolocation databases. The great majority of caching DNS servers are able to essentially act as proxies for IPv6-only clients to do lookups.--Jasper Deng (talk) 02:37, 8 March 2020 (UTC)

@Jasper Deng: Ah, that makes sense. Thank you for the explanation. - Dyork (talk) 03:08, 8 March 2020 (UTC)

Someone said ...

In 1970-s, when IPv4 was developed, the population of Earth was 4 billions (it became twice more during previous 50 years). Developers of IPv4 must understand that 4 billions of addresses will not be enough for every host on the planet. Short 32-bit address was chosen with intention of future implementation of NAT and private networks. Short 32-bit address makes NAT and private networks necessary. That is the goal of short 32-bit address. Long 128-bit address eliminates this. Gah4 (talk) 21:59, 30 March 2020 (UTC)

I believe the above is not historically correct. When IP was developed, computers were expensive. I suppose the designers should have seen what might happen, and planned ahead for it, but they didn't. At that time, many machines used (at least) 32 bit words, or some used 16 bit words but could do 32 bit arithmetic in two operations. It was a convenient size. One example of how much they didn't plan is allocating a whole class A network, 127/8, for the loopback address. Initially, there were only three classes, with /8, /16, and /24 nets, not quite optimal for the existing sized organizations. Gah4 (talk) 22:06, 30 March 2020 (UTC)

So, why did you remove entire chapter? Why not this paragraph only? Why did you remove IPv6 image? May I come back another paragraphs? If you don't justify removing. Acbaile5 (talk) 22:10, 30 March 2020 (UTC)

I will explain the same more academically. OK? Acbaile5 (talk) 22:15, 30 March 2020 (UTC)

I copied that one, as I think with some changes it could be used. NAT and private networks are a different question. Private networks first came when people (usually companies) had networks not connected, and not planned to connect, to the internet (rest of the world). They still needed IP addresses, so some were assigned. At some point, the idea of NAT, firewall, and private networks came together to satisfy keeping networks safe, and the shortage of public IP addresses, at about the same time. Even with IPv6, it is probably best to have a firewall to allow only appropriate traffic through. Gah4 (talk) 22:30, 30 March 2020 (UTC)

When IPv4 started in 1983, very few computers were on the net. It spread fairly fast to academic institutions, but even so maybe one or two hosts per department, mostly for mail. (Previously, mail was done through UUCP and dial-up phone lines.) Early on, much of the internet was 64k bit/s links, good for mail, and small file transfer. PCs with 16K RAM were popular, but not much thought to connecting them to the world, though maybe to a small local (small i) internet. (Now usually named intranet.) If you want to research the development of IPv4, ARPAnet, NSFnet, through to the public (not government run) Internet, and then to IPv6, that would be nice. Gah4 (talk) 22:30, 30 March 2020 (UTC)

Oh, I found this one indicating what one might call a computer in 1983. But more usual were VAX and such. Gah4 (talk) 22:30, 30 March 2020 (UTC)

Word "NAT" has sound similar to "NOT!!" with two exclamations. It looks like the sense of this technology: the men don't accept totally opened network. So, private networks must exist, and must be transparent inside particular network only. Acbaile5 (talk) 22:38, 30 March 2020 (UTC)

I added first paragraph that touches possible IPv6 danger only. I tried to write academically. Acbaile5 (talk) 22:34, 30 March 2020 (UTC)

Some centuries ago cheap books appeared. Now we don't burn these books like Library of Alexandria. Men simply publish a lot of fiction books, so it becomes really hard to find some book with true. Doesn't Google resolve analogic problem with Internet?

Why do I like global Internet? At first, because of global safety. Transparent world is much more safer then a set of closed communities. I don't like Ares, I don't like war. I understand that men need control and usually like blindness, but absence of war is principal moment. I accept existence of lying Google. Every smartphone is a PRESS on your mind - simply open any web page - and a stream of lie will come into your eyes, where ever you are. Isn't it using of Internet for men benefits? Now you want simply press without safety? Acbaile5 (talk) 23:32, 30 March 2020 (UTC)

@Acbaile5: You can write all you want about the internet being a "press" but unless and until you demonstrate the existence of reliable sources espousing this exact view, it's not suitable for Wikipedia. I strongly agree with the removal of both of your additions because besides the numerous grammar errors in them, they are woefully incorrect, and only go to show how ill-informed you are (this is why we must use reliable sources. Even someone with a degree in computer science like me cannot substitute his opinions for what reliable sources have written). For starters, NAT has not proven to be a deterrent to the assembly of global botnets and that problem will not be made worse by the removal of NAT. NAT is also not a "social" thing (that is absolutely not the meaning of a private network, nor will IPv6 mean the complete end of the use of local addresses). There's also no evidence whatsoever to suggest IPv6 will be the end of the client-server model (except perhaps for things explicitly intended for NAT traversal like Hamachi (software)). We are not here to WP:RIGHTGREATWRONGS and per WP:NOTAFORUM, please don't pursue this any further unless and until you provide reliable sources supporting your view.--Jasper Deng (talk) 23:46, 30 March 2020 (UTC)

NAT was convenient in that it provides some isolation without much complication in configuring, especially for the usual home network with only outbound connections. Except that some home uses, such as VoIP need to cross the NAT. However, a firewall is a much better solution than NAT. Personally, I run a unix-like machine with three ports, for two house subnets, including IPv6 routing. I don't know what more usual home routers do as far as IPv6. The use of private (not connected to the Internet) IPv6 networks would be a useful addition to the article. Gah4 (talk) 01:21, 31 March 2020 (UTC)

I don't think many folks are actually making use of the private IPv6 addresses. I know for sure that the WMF is using public IPv6 addresses for servers with private IPv4 addresses. Leslie Carr once told me that IPv6 can be just as secure as IPv4 NAT with appropriate firewalling. But it's been a while since I've checked out books on this subject. There might be some coverage now.--Jasper Deng (talk) 02:41, 31 March 2020 (UTC)

Well the might be, but not telling anyone about them. That is the whole advantage of a private net. In IPv4 days, I was working with someone installing our (expensive) product in some company. One part of installation is sending some configuration data outside. It seems that this company had no way to get e-mail out, so it had to be faxed. (Well, maybe 20 years ago.) It is said that election processing systems are not connected to the Internet, though likely internally networked. But then again, IPv4 is probably fine. Gah4 (talk) 05:25, 31 March 2020 (UTC)

The only way to find out is to see if sources report on it, though. A good resource would be a handbook like the For Dummies series on practical networking.--Jasper Deng (talk) 06:27, 31 March 2020 (UTC)

@Jasper Deng: No, it looks like intention to hide the potential danger of IPv6 from Wiki readers. You simply don't want that people see any idea about it, that's why you eliminate it from Wikipedia. It is not justifying, it is censorship.

Grammar errors are present because English is not my native language. You may correct grammar errors. But I don't see technical errors.

"woefully incorrect", "ill-informed" — :) Sorry, it is not technical justifying. It looks like eliminating of inconvenient thoughts, far away from eyes.

Removing of NAT is distant target of IPv6. We will not see it, but our descendants will see it. IPv6 is intentional building of broken bridge. It will collapse. When a lot of users will have absolutely freedom of communication in IPv6 - it will collapse. Private networks and NATs in IPv4 deter absolutely freedom of communication. But nothing will deter this freedom if everybody will have globally routed address. Long 128-bit address of IPv6 is the goal of this absolutely freedom. The problem has social nature. Men want to control. Absolutely freedom of communications in IPv6 removes the possibility of control. Consequence: men are angry, men want to destroy IPv6. IPv6 provokes this conflict. This conflict will appear in future. IPv6 is intentional weapon of this conflict. IPv6 is built for it.

IPv4 gives possibilities of control. But, it looks, that is not enough for you. It looks you want to eliminate possibility to say opinion free in Internet. It seems, it's too much freedom for you.

Server is necessary when private networks exist, when clients can not establish connections directly. Server is one point where user's activities are concentrated. It is much more simpler to watch and control one certain point instead of any number of unpredictable points in absolutely transparent network. Server becomes unnecessary when clients can establish connections directly. Men will not be able to watch and control these direct connections. That is the reason why men will want to destroy this freedom of communications.

Acbaile5 (talk) 13:20, 31 March 2020 (UTC)

@Gah4: "NAT was convenient..." — That does not explain why NAT does not give possibility to know which port number is assigned to your request when translation takes place. It's done to eliminate the possibility to establish incoming connections to the clients in private networks. It must stay. Private networks must stay semi-closed.

"However, a firewall is a much better solution than NAT" — You offer the same freedom. Of course, firewalls will exist in IPv6, network can not be fully opened. But users will be able to open any necessary port. That will give absolutely freedom of communications that NAT deters.

Acbaile5 (talk) 13:31, 31 March 2020 (UTC)

@Jasper Deng:"... IPv6 can be just as secure as IPv4 NAT with appropriate firewalling" — Yes, of course, it can be, but it will not be :) . Because of the opposite intention. Appearing of IPv6 is not accident. IPv6 is intended for destroying the globality of Internet. So, I don't believe that IPv6 will be used accurately. Generations will change, and our descendants will see collapse of global network.

You still did not explain why you removed nice image of IPv6 like time bomb. It seems you understand what this picture means.

It seems, we need some community against IPv6 - to have right to imprint the danger of IPv6 into the wiki, for our descendants.

Acbaile5 (talk) 13:44, 31 March 2020 (UTC)

@Jasper Deng:"A good resource would be a handbook like the For Dummies series on practical networking" — Can be the solution making Google-like systems more efficient? Or it will not be enough? Acbaile5 (talk) 14:30, 31 March 2020 (UTC)

@Acbaile5: The only way you can convince us to add anything of this sort is if you find reliable sources espousing this view, period. You clearly do not at all understand how computer networking works because users cannot open ports that their ISP or organization blocks! Many ISP's block the standard SMTP ports for this reason. And also, the server which the user wants to reach always knows the port the user is trying to connect to it on (how else can it serve the user otherwise?). The existence of NAT traversal techniques, which are sometimes peer-to-peer (sometimes even as simple as port forwarding, which I have used to great effect in the past), as well as peer-to-peer protocols like BitTorrent, is constructive disproof of your notion that the "server is necessary". Again, too, this talk page is WP:NOTAFORUM; you are wasting your time and energy here without providing sources for your claims.--Jasper Deng (talk) 19:41, 31 March 2020 (UTC)

OK, the reason I started this section is that I believe something might be said in the article related to the address space sizes (of Iv4 and IPv6), the reasons for those, or faulty reasoning in those, choices. As noted, 32 bits isn't near enough, why didn't they know that in 1981? (1983 is deployment, decisions were made earlier.) Why do we have NAT? (Security or addressing?) What is a good way to secure home networks? (NAT or firewall?) It seems so obvious now that everyone in the world should be on one Internet, but that wasn't at all obvious in 1981. Gah4 (talk) 17:44, 31 March 2020 (UTC)

Vint Cerf has apologized for the 32-bit decision multiple times, linked from this article. Ttwaring (talk) 18:34, 31 March 2020 (UTC)

The question, though, is should that be in this or IPv4? I suspect,though, that it made sense at the time. There is much written (somewhere) about the decision to make the 16 bit PDP-11, for which 16 bit addressing was not enough almost immediately. The 32 bits for VAX lasted quite a while, though it doesn't allow 32 bits for a single task. At that point, though, Ethernet already had its 48 bit address, so that might have been a hint. There really weren't that many computers at the time that anyone would want to network. Gah4 (talk) 19:16, 31 March 2020 (UTC)

@Gah4: I would think NAT was not originally conceived for address exhaustion mitigation, so I'm curious too as to the original reasons for inventing it. But what I can tell is, IPv4's 32-bit address was in part due to the perception that the Internet would not be anywhere near its present size, hence why some 64-bit address ideas (IIRC) were shot down. We need reliable sources either way.--Jasper Deng (talk) 19:41, 31 March 2020 (UTC)

@Gah4: another place this can go is IPv4 address exhaustion ~Kvng (talk) 13:02, 3 April 2020 (UTC)

As well as I remember, NAT was originally when some ISP would not issue whole subnets, especially for home use. Partly that was because they thought there was one customer, not a whole house full. (Yes, a long time ago now.) Many years ago, I had my home network linked to the outside world through a 9600 baud modem permanently dialed up. At one time (about 1992), I had a real subnet with PCroute for routing, but then moved to a different place (about 1995) where it didn't, and I used, if I remember, IProute, PC based routing software implementing NAT. Whether or not there were enough addresses, ISPs weren't issuing whole subnets. Not so much later, the place I was working had DSL for networking. At special request, we got a /29 subnet, so six addresses. For that, I used FreeBSD for a NAT router running on an otherwise unused 80486 box. Exactly when it was realized that IP addresses would run out isn't so obvious, but ISPs weren't issuing /24 nets for home use, even if they could. At some point, there was a transition from nerds with home networks to ordinary people with internet connected home computers, and then whole home networks. I presume that there are WP:RS for this, but I haven't looked. Gah4 (talk) 19:31, 3 April 2020 (UTC)

Future of network

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

Let's discuss about the future of network - to find solution that will be acceptable for both sides. Let's imagine that IPv6 came to life, private networks disappeared, chaos in using happened and governments decided to separate Internet on small segments. What form will it take and what are traits desirable for men in future network? Maybe, we can find some solution without total destroying?

For example, my offer is implement the concept of sovereignty over the informational space for the state in international law. That means the state has right to restrict access to information.

For women, it is very desirable to keep some part of globality - particularly, access to news sites, because it is important for global safety. It's important to see what happens inside other countries. Is it possible to keep it?

Another reason is the mail. But the mail, in some form, will exist ever. Even if network will be a set of intranets, the mail will find a way to communicate. Global IPv4 makes it easier, it's true. But safety is more important than mail. Let states have right to restrict access to the mail. Now, the mail can be given to other nations by mistake, because of the lie. People that want war make this lie to provoke war. I mean they say: "All nations are equal, just make this gift for ALL nations". Right to restrict will eliminate this problem - any state will have possibility to restrict at any moment.

Maybe, you want more hierarchical net? Net of city is more transparent in borders of city, net of state is less transparent, international net is less transparent than net of state. Let's implement it without chaos. That is the social nature of private networks.

Acbaile5 (talk) 16:02, 31 March 2020 (UTC)

No, let us not discuss, because this has nothing to do with improving the article. @Acbaile5: As this section is clearly in contravention of WP:NOTAFORUM, I am closing this. We are not here to publish your original research.--Jasper Deng (talk) 19:43, 31 March 2020 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

• Comment - original research?? That's putting it mildly, to say the least :P. Meseemeth it more like they mistook this talk page for the United Nations. Firejuggler86 (talk) 01:10, 11 April 2021 (UTC)

squares

The article says: square of the size of the entire IPv4 address space. I suppose this is true, but address space is most often discussed in bits, that is, the logarithm of the number of addresses. It is also complicated by the use of units. If I have 10 apples and square them, then I will have 100 square apples. (More obvious, 10cm squared is 100 square centimeters, a convenient area unit.) I suspect that people who don't understand bits and logs also won't understand a square address space. Gah4 (talk) 17:16, 18 August 2020 (UTC)

Agreed, I've simplified this to remove the "square" part. Ttwaring (talk) 17:40, 18 August 2020 (UTC)

Right date

We should check the date November 1998 as when IPv6 became a standard! Vladimir Skokan1 (talk) 20:22, 26 January 2022 (UTC)

I meant December 1998 in EATF! Vladimir Skokan1 (talk) 20:24, 26 January 2022 (UTC)

I know it's been over a year, but when I read the above acronym "…in EATF!" I was confused… I did figure out eventually that it was meant as IETF — the article itself, though, has the correct term. RobiBuecheler (talk) 19:55, 21 May 2023 (UTC)

No Criticism of IPv6?

One of the most controversial protocols in the history of computing and there's no controversy or criticism section...?

I guess IPv6 is perfect. My mistake. — Preceding unsigned comment added by 76.105.169.30 (talk) 03:40, 3 June 2022 (UTC)

It's probably not perfect. The first step here would be to suggest reliable sources with criticisms. Ttwaring (talk) 14:32, 3 June 2022 (UTC)

I've yet to hear criticisms from anyone with any amount credibility on the subject. 2605:4A80:B401:AE41:5144:2078:4A08:AEF0 (talk) 14:29, 14 July 2024 (UTC)

OK, this might be WP:OR, but then again, that is allowed in talk. I haven't found a WP:RS yet.

My home network is properly configure with IPv4 addresses, and forward/reverse DNS records.

But it seems that IPv6 hosts keep changing their address. The network part stays, as it is supposed to, but the host part doesn't!

It seems to be part of some security system, and I did look for some WP:RS to see what others do about it.

I haven't found one yet. Gah4 (talk) 20:51, 15 July 2024 (UTC)

It's from RFC 4941, "Privacy Extensions for Stateless Address Autoconfiguration in IPv6". See IPv6#Stateless_address_autoconfiguration_(SLAAC). Ttwaring (talk) 21:42, 15 July 2024 (UTC)

OK, so I am trying to criticize it. For my home network, I have configured 100 addresses for DHCP, and put DNS entries for them. All works fine. But I can't put 1.844674407370955e19 DNS entries that DHCP6 could assign. OK, I did know it was SLAAC, but reading that one did help. A Google search finds: In its current implementation as defined in RFC 4862, SLAAC does not provide DNS server addresses to hosts and that is why it is not widely adopted at the moment. It seems that it a criticism. Gah4 (talk) 07:35, 16 July 2024 (UTC)

SLAAC can provide DNS information using RDNSS (RFC 8106) as part of Router Advertisements now. 24.230.161.142 (talk) 13:22, 16 July 2024 (UTC)

If you want static hostid, you can configure that. It is not a criticism of IPv6 though, and this page is not an IPv6 configuration help forum.

I would oppose a criticism section per WP:CRIT, but we might want to mention some of the contentious decisions that were made and why. The most obvious one was the decision to go with 128 bit addressing instead of 64 bits. Arguments were made (and they were right) that 64 bits were sufficient that no further expansion would ever be required. 128 bits was overkill. But the decision to go with 128 bits was precisely because of considerations of autoconfiguration and the like. This added 128 bits (2 x the extra 64) to every single datagram though, which was not insignificant. Now that is in some books, and I could dig out a reference if we wanted to cover it. Sirfurboy🏄 (talk) 13:44, 16 July 2024 (UTC)

Problem is lack of reliable sources. Criticism of IPv6 is sort of career suicide for network engineers. IETF is not going to criticize themselves, instead their focus is on explaining why they made the choices they made, and on writing more specifications that would helping transition. IPv6 was the 'best effort' at the time, so nobody is really sure what choices could have been better (because the alternative does not exist). In my view very much a missed chance, instead of "lets aim for good-enough and just expand the address space in a compatible way" it was "lets rewrite the protocol and make it better" which explains the consequences of an extremely slow rollout. The term Worse is better captures this sentiment pretty accurately. But unlike the software world in this case there are no competing alternatives. Sdesalas. (talk) 11:45, 4 September 2024 (UTC)

It was impossible to just "expand the address space in a compatible way" because IPv4 is and always will be limited to 32-bit addresses. This is why IPv6 pushed forward the way that it did. 2605:4A80:B401:AE41:6168:24D8:385F:501C (talk) 00:51, 15 September 2024 (UTC)

I did find this source where the Internet Society actually points out that the lack of backwards compatibility is indeed the biggest flaw in IPv6. However, the source is a bit dated, and I don't think it adds much new information beyond what is already covered in the article. Regarding your point, I would say that hindsight is always 20/20. It's similar to how some people argue that network designers defining IPv4 in the 1980s should have predicted the ubiquity of modern-day internet devices. IPv6 development began in the mid-90s, and at that time, replacing IPv4 with IPv6 may have seemed reasonable given the size of the Internet and the number of connected devices - had they predicted the size the internet would grow to before the standard was even completely defined, they may have considered otherwise. askeuhd (talk) 19:59, 22 September 2024 (UTC)

Rename

There is discussion in Talk:Internet Protocol version 4 on renaming to IPv4. One of the reasons for the rename, is to agree with this name. Seems to me, though, that could equally be used to rename this one. Until IPv6 came along, there was no need to call the previous one IPv4 or Internet Protocol version 4 as it was the only version ever used outside the development labs. (It would be nice to know about 1, 2, 3, and 5, though.) Gah4 (talk) 19:46, 5 May 2024 (UTC)

List of IP version numbers is a good starting point for versions 1, 2, 3, 5, and more. Ttwaring (talk) 15:59, 7 May 2024 (UTC)