The only case where I care about an IPv6 address is for something I actually want to expose to the internet. A temporary address would be quite annoying in that case.
It's fdac::1. If you're using random 48 bit or 64 bit numbers in your IP address you're doing it wrong.
I have zero concerns that the IPv6 namespace for my home network will conflict with another administrative site during a merger. So.. it works great. Also super handy when the DNS resolver for my local network is down because of power outages or other unrelated failures.
A home network has neither of those, you're running a nerd network. You can set up an mDNS relay, and those devices can use other methods to register their addresses & get DNS updated (DHCPv6, DDNS, random cloud services, etc.).
I think the Matter standard is going to cause IPv6 adoption to increase significantly in the coming years. People will demand it, without even knowing what Matter or IPv6 are. They just want to be able to turn their lights off from their phone without any extra hardware or software.
Matter is only local networking. Essentially everyone has IPv6 on their local network.
That said, I disagree with the parent statement that v6 isn't going to happen. It's already happening at a steady rate. From the Google stats it's steadily rolling out. Some countries are reaching 100% v6 deployment. I wouldn't be surprised if we eventually see some countries with high v6 deployments see services which are v6 only.
They've had years to demand it, I don't see why it would suddenly change. Most users who buy "smart" appliances are not power users, don't set everything up themselves and just rely on manufacturer's servers anyway and don't need a direct connection - and are okay with it.
There’s something to be said for human readable addresses. I’m a little nostalgic of how the .hack world was envisioned, where servers had address names like Hidden Forbidden Holy Ground.
If roughly 10 million words exist, then allowing any three words in order creates a space for 10^21 addresses… five words and you’re close to ipv6 address space, six words and there’s more combinations than ipv6 addresses.
You can set your ULA to something like "fddd:192:168::/48" and then on your vlan you prefix hint, say, "66". Now, any device on that vlan will be addressable by "fddd:192:168:66::$host". For example, your gateway ('router') for that vlan would be "fddd:192:168:66::1".
If you want to be really wonky you can script DHCPv6 to statically assign ULA IPv6 leases that match the IPv4, and expire them when the IPv4 lease expires, but like said upthread, addressing hosts via IPv6 is the wrong way to go about it. On your lan, you really want to be doing ".local" / ".lan" / ".home".
> addressing hosts via IPv6 is the wrong way to go about it. On your lan, you really want to be doing ".local" / ".lan" / ".home".
.local is fine as long as all the daemons work correctly, but AFAIK there's no way to have SLAAC and put hosts in "normal" internal DNS, so .lan/.home/.internal are probably out.
> On your lan, you really want to be doing ".local" / ".lan" / ".home".
The "official" is home.arpa according to RFC 8375 [1]:
Users and devices within a home network (hereafter referred to as
"homenet") require devices and services to be identified by names
that are unique within the boundaries of the homenet [RFC7368]. The
naming mechanism needs to function without configuration from the
user. While it may be possible for a name to be delegated by an ISP,
homenets must also function in the absence of such a delegation.
This document reserves the name 'home.arpa.' to serve as the default
name for this purpose, with a scope limited to each individual
homenet.
It may be the most officially-recommended for home use, but .internal is also officially endorsed for "private-use applications" (deciding the semantics of these is left as an exercise to the reader): https://en.wikipedia.org/wiki/.internal
".home" and ".lan" along with a bunch of other historic tlds are on the reserved list and cannot be registered.
Call techy people pathologically lazy but no one is going to switch to typing ".home.arpa" or ".internal". They should have stuck with the original proposal of making ".home" official, instead of sticking ".arpa" behind it. That immediately doomed the RFC.
I do it by abusing the static slaac address. I have a set of wierd vms where they are cloned from a reference image, so no fixed config allowed. I should have probably just have used dhcp6 but I started by trying slaac and the static address were stable enough for my purposes so it stuck.
How does that work? I initially assumed you meant you just statically assigned machines to addresses, which I think would work courtesy of collision avoidance (and the massive address space), but I can't see how that would work for VMs. Are you just letting VMs pick an IP at random and then having them never change it, at which point you manually add them to DNS?
Pretty much. A given mac address assigned in the vm config maps directly to a static slaac address(the ones they recommend you not use) and those preknown slaac address are in dns, Like I said, I should probably use dhcp6 but it was a personal experiment in cloning a vm for a sandbox execution environment. and those slacc address were stable enough for that. every time it gets cloned to the same mac address it ended up with the same ip6 address. works for me, don't have to faf around with dhcp6, put it in dns. time for a drink.
But the point is that is the address you would put in dns if you also wanted to use slaac. Most of the time however you will just set a manual address. And this was with obsd, where when slaac is setup you get the slaac address and a temporary address. I don't really know what linux does. Might have to try now.
Clarification for others: with privacy extensions disabled, SLAAC'd IPv6 addresses are deterministically generated based on MAC addresses. There's also an inbetween (IPv6 are stable per network by hashing).
You have to take into account seasonal trends. The summer is always higher, so yes, we’re currently below last summer, but we are above last April 1st, and this summer will be higher than last summer.
You can’t compare fall winter and spring to last July. You have to compare last July to this July, which hasn’t happened yet, but when it does, it will be higher than last July. Today (April) is higher than last April.
The reason for this is that more people are on mobile connections during the summer (kids home from school) so the summer (as well as the Christmas/new year week) are the highest points of the year.
The fact that it’s “flat” since last July, the high season, means it’s actually still increasing.
> Regardless, when you extrapolate that curve, when do you estimate hitting 100%?
Never? But what’s your point? IPv6 is a failure if it only replaces IPv4 for 99% of traffic?
I’m always bemused when I read comments like this. Regardless of whether you think IPv6 is good or bad, it is happening. IPv6 traffic to Google goes up by a few percentage points a year, steadily, and is at around 50% now.
They should charge a small annual fee and let people reserve a custom word for a given IP. You could even have a small utility on your computer that automatically queries given names to "resolve" to IPs.
To facilitate client to client communications, they should add a who-does-it-serve (short WHO DIS) system to get contact information of already reserved names.
And then if they don't like someone they should revoke their word. And they should make it so words with certain endings are delegated to other countries who can also revoke your word if they don't like you.
I tried it. Maybe it's easier to speak than hexadecimal is.
But I'm not sure that "How morally the enviable assistances categorize the insistent iodine beyond new time where new systems stalk" has the same memorable quality as "correct horse battery staple" does.
Being essentially impossible to memorize is one of the worst attributes of IPv6. I memorize and manually type IPv4 addresses all the time and it's super useful.
> My DNS "server" is a router which can "add" static entries...won't work with dynamic addresses.
Sounds like a pretty poor setup, systems which could auto-add DHCP'd or discovered entries have been around for literally decades. You're choosing to live in that limitation.
> What redundancy, multiple servers?
Multicast name resolution is a thing. Hosts can send out queries and other devices can respond back. You don't need a centralized DNS server to have functional DNS.
Oh yes, that's really convenient for home users. "Install this thing on several computers and keep it in sync or you're not qualified to have a network"
Home users would ideally be served by things like mDNS and LLMNR, which should just work in the background. If I want to connect to the thermostat I should be able to just go to http://honeywell-thermostat and have it work. If I want to connect to the printer it should just be ipp://brother and I shouldn't even need to have a DNS server.
Your interface has a default gateway configured for it, doesn't it? Isn't that default gateway the router? NDP should show the local routers through router advertisements. There is also LLDP to help find such devices. LLMNR/mDNS provides DNS services even without a centralized nameserver (hence the whole "I shouldn't even need to have a DNS server"). So much out there other than just memorizing numbers. I've been working with IPv6 for nearly 20 years and I've never had an issue of "what was the IP address of the local router", because there's so many ways to find devices.
Even then nobody is stopping you from giving them memorable IP addresses. Giving your local router a link-local address of fe80::1 is perfectly valid. Or if you're needing larger networking than just link-local and have memorable addresses use ULAs and have the router on network one be fd00:1::1, the router on network two be fd00:2::1, the router on network three be fd00:3::1, etc. Is fe80::1 or fd00:1::1 really that much harder to memorize than 192.168.0.1 or 192.168.1.1 or 10.0.0.1, if you're really super gung-ho about memorizing numbers?
Adding two extra bits to each octet, making each octet range from a still memorable 0-1023 rather than 0-255, would result in an addressing scheme 256x larger than all of IPv4 combined. The entire internet works fine even when IPv4 was nominally exhausted. NAT and CGNAT are not sins, they're not crimes, and there's no rational reason to be as disgusted with them as IPv6 fans are. Even then, IPv4 exhaustion wasn't really a true technical problem in the first place, it was an allocation problem. There are huge /8 blocks of public IPv4 space that remain almost entirely unused to this day.
The reason I'm an IPv4 advocate in the IPv4/IPv6 war is that the problem was "we're out of address", not "your thermostat should be natively routable from every single smartphone on the planet by default and inbound firewalls should become everyone's responsibility to configure for every device they own".
CGNAT is a feature, not a bug. Blending in with the crowd with a dynamic WAN IP is a helpful boost to privacy, even if not a one-stop solution. IPv6 giving everyone a globally unique, stable address by default is a regression in everyone's default privacy, and effectively a death sentence for the privacy of non-technical users who aren't capable of configuring privacy extensions. It's a wet dream for shady data brokers, intelligence agencies, organized crime, and script kiddies alike - all adversaries / attackers in threat modelling scenarios.
IPv6 adds configuration surface I don't want. Privacy extensions, temporary addresses, RA flags, NDP, DHCPv6 vs SLAAC — these are problems I don't have with IPv4. More features means more opportunities to footgun with misconfigurations, being forced to waste my time learning and understanding the nuances of each (in again, what amounts to system I want nothing to do with).
"Reaching your own stuff" is already a solved problem, too. Tailscale/Headscale gives you authenticated, encrypted, NAT-traversing connectivity. It's better than being globally routable. It's also opt-in for anyone who wants it, and not forced on anyone, unlike the IPv6 transition.
I don't have your problems with ipv6, and I'm actively using it.
I don't have to rely on extra commercial entities to be able to reach my network.
I did have a problem with hosting my own shit because my ISP by default does cgnat. That cost me an hour of my life to convince a party to give what used to be normal, end to end connectivity.
yes you do, the control plane is closed. Only reverse engineered by the headscale project. The control plane is necessary for the peers to find each other. If you need to rely on such a crucial part being reverse-engineered, than yes, I think it's fair to say you are ultimately relying on commercial entities.
Wouldn't easy and accessible self-hosting be a major privacy win if that's your primary concern? Sounds much more private to run a Minecraft and Mumble server on an old laptop in a friend group than paying a commercial entity like a hosting provider to know about it and have a back door.
Easy and accessible self hosting isn't the primary concern.
It's much more private and secure to run that Minecraft or Mumble server on an encrypted overlay network like via headscale + tailscale rather than exposing both services directly to the entire planet.
But again, the primary concern was only ever address space.
What I tried to express was privacy being the primary concern. The easy and accessible self-hosting on old hardware would be the uses of a home network beyond superficialities like consumption and commerce. Privacy wise headscale as a solution is still not quite there, because it either necessitates an additional third party to host the headscale server and know about all my friends, or jank like dynDNS.
The additional security gained by getting everyone involved to set up and configure separate VPNs for different community utilities is not worth it.
I wouldn't call port forwarding "huge problems". It's only one minor router setting and if you don't want to deal with it, there's the abomination called upnp.
> I wouldn't call port forwarding "huge problems".
Port forwarding has massive problems if you're running applications expecting certain ports and need multiple hosts to have public access to those ports.
Highly disagree. Middleboxes are a huge problem on global scale and have frozen any innovation below application layer. TCP and UDP even that they are on software not hardware layer cannot be updated or changed, see MPTCP efforts or QUIC giving up and building on top of UDP.
If this is so much privacy problem, IPv6 is there for many years reaching 50%+ deployments in some countries, I bet there should be concrete examples of such breaches and papers written.
> Reaching your own stuff is already a solved problem, too. Tailscale/Headscale
No address to receive communication - no problem install an app that would proxy it through someone who has the address.
Tailscale/Headscale is great, using it daily, but they are not solution to the huge already build global network created to connect devices not connecting devices because lack of digits. Global is key here.
Kind of like what3words, except what3words uses three words which you stand a chance of remembering, whereas this produced, for an address similar to mine, "Miniature nerves eulogize gaily inside erect lion yet able stables hiss the conclusive consultation."
Care to elaborate why? I'm just curious since I didn't know (1) there was actually any kind of serious usage, or (2) there was pushback from rescue team...
W3W is very aggressive about protecting their IP, they don't want it to be a standard anyone can use like lat/long.
They advertise it as being useful for search/rescue as you can provide a precise location over an unclear voice channel. They conveniently ignore that speaking numbers is clearer than speaking random words.
One reason is that there are a lot of similar words in the dictionary. It is easy to mishear the wrong location especially when they are close together. Some of the words are long and complicated. Another is that they are random which means can't navigate from the codes.
They are missing feature of some codes that can have variable length for variable precision.
Something that I think was probably once obvious to me but I rediscovered recently is just how intensely wired for song the brain is. If you want to memorize anything, doing it as a song makes it far easier.
I’d really love to see things like this generate little jingles along with the sentence. :)
There actually was an attempt on HN a little while back to use GenAI to convert facts, flashcards, lists, etc. into automated melodic mnemonics. The biggest issue in that particular case was that it was also generating the motif from scratch.
At least for me, part of the reason I can still sing the countries of the world is because the original Animaniacs song was set to a tune that was already familiar: “Jarabe Tapatío” (aka the Mexican Hat Dance).
I memorized that (and several other) Animaniacs songs without being familiar with the melody. Even Tom Lehrer’s The Elements reached me before Pirates of Penzance did. I think the melody just needs to be simple, then it'll become ”familiar” quickly.
However, for the use-case at hand (remembering IPv6 addresses) I don't think I'd use that. I'd just write them down somewhere, like, uh, perhaps, oh I know: the hosts file.
“Catchiness” is probably more important than anything, hence the concept of the earworm aka stuck song syndrome. Even SOTA GenAI like Suno/Udio fall pretty short of generating genuinely engaging melodies.
As anecdotal evidence, I could never remember all the names of the london tube stations as prose, but after listening to Jay Foreman's Every Tube Station song a few times, I cant help but sing it every time I hear a tube station name, and can almost flawlessly list them all off
What is the use-case for this? I’m trying to think of an IPv6 address I would need to remember, and then when I’d have access to this site without having access to a text file where I could have noted the address down. I’m coming up empty.
It'd be nice if simpler addresses gave simpler sentences. fe80::1 translates to "Uninhibited times take now inside new time yet new times take the new year."
I don't understand how the mapping works. An address has 8 parts and produces 16 words, so each part consists of 2 words. If we take the example 2a02, that gets encoded to "how atop", but I don't see how that text helps me that "how atop" means 2a02? Am I suppose to memorize both? How does that help?
You are not supposed worry about the mapping. You trust the website to help decode it. You just remember the sentence. It's a little like what3words for coordinates.
The rationale being you are more likely to remember grammatical cogent sentence, than a random string of alphanumeric characters. Although I will agree that the generated sentences don't seem easy to remember. So I doubt it's utility.
So just imagine if there was a service that could translate any words you wanted into the IP address instead of relying on some website to generate jibberish. Wouldn't that be cool to use instead? Some kind of name system? Based around domains of authority?
Now, if only those people who designed IPv6 were smarter.. Hex aint that bad, LONG hex addresses are pain to use.
Now, lets say you have LAN like this [::1:0:0/56]. So, ::1:0:24 is easy to remember right? Managable? right?.. Also, bonus for :: shortening is, you immediatly know what are you dealing with, ::1 is loopback, ::1:1 is LL, ::1:0:1 is LAN.. everything else is Internet.
The truth is, IPv6 is really 64bit, the other 64bit part is just randomish node address...
> How now the smart flies take the new time beyond new time where new times come.
..Nice idea, but it may need some more thought. (Even more so as 2001:db8::1 is much easier to remember than that!) (I wrote that parenthetical from memory on edit, vs. had to copy-paste the sentence when it was my intention to comment on it within seconds.)
Not too sure of the utility of this. It's not an easy sentence to remember, because while grammatical, it's nonsense—it would take some effort. So if I'm trying to memorize a static IP, setting up a DNS name is likely to be easier. And also if I'm going to be using this to memorize IPs I'd like the algorithm to be open source.
All that being said, I think it's a neat idea and a cool tool!
We kind of had the same idea for ECDSA public keys (an imagined solution to zokos triangle -- human readable and decentralized) as well as private keys (BIP39 brain wallets). Honestly it still falls short of truly name-based though.
I'm fine with 16 but they should have only used the bytes as they were needed, at least for 5 and 6 byte addresses, so those who desire short addresses could buy them.
fe80:: is for link local. You'd want to use something starting fc00:: or fd00::
In your typical home environment, just set your ULA to fd00::12 instead of 192.168.0.12, or fd00:16:34 instead of 192.168.16.34
Yes you'll run into issues if you were to later want to merge your private IPs with someone else, and you should use fd12:3456:7890::12 instead, remembering those extra 10 digits, but its not a problem at home, and no more of a problem with business mergers than ipv4 clashes anyway.
My vrrp address for my dns server at home is 2001:8b0:abcd::53
It's about as easy to remember as 81.187.123.45//192.168.0.53
Almost all ipv6 addresses I encounter start with 2001, so I just need to remember my home prefix is 8b0:abcd, which is about the same length as my home public IP of 81.187.123.45
::53 means subnet zero host 53, which is easier to remember than which rfc1918 range I used (and basically is the equivalent of remembering the 2001:: prefix)
If I have an internal server I'd have on 192.168.4.12 I could address it with 2001:8b0:abcd:4::12 just as easily, with the "4.12" translating to "4::12", and the "81.187.123.45>192.168.x.y" translating to "2001:8b0:abcd:x::y"
Just because slacc gives you an extra 64 bits of randomness doesn't mean you need to use them.
If you're assigning addresses, you can make the addresses in a ULA as short as you want. You're supposed to use a random 40 bit network id but if you can accept that you may need to renumber at some point there is no reason you can't use fd12:b:a:d::beef or whatever.
How could we determine which device on mobile network is a faceless cellphone and which is a proper device needing real sweet Internet connection? And won't that make things more complicated than just v6 deployment?
Can argue that NAT, which interrupt layers ment for end device do basically the same as popular user hostinle unchangable mobile OSes, but I don't think latter is good either.
If you're remembering your IPv6 address you're doing IPv6 wrong. In fact, it's good practice to always use a temporary IPv6 address.
https://datatracker.ietf.org/doc/html/rfc8981#name-problem-s...
The only case where I care about an IPv6 address is for something I actually want to expose to the internet. A temporary address would be quite annoying in that case.
It's fdac::1. If you're using random 48 bit or 64 bit numbers in your IP address you're doing it wrong.
I have zero concerns that the IPv6 namespace for my home network will conflict with another administrative site during a merger. So.. it works great. Also super handy when the DNS resolver for my local network is down because of power outages or other unrelated failures.
... How do you connect to another PC on your network? Always use DNS or something?
> Always use DNS or something?
Yes. mDNS if we're talking about home networks.
That doesn't work across VLANs very well, how do you manage that?
Or devices that don't offer mDNS?
A home network has neither of those, you're running a nerd network. You can set up an mDNS relay, and those devices can use other methods to register their addresses & get DNS updated (DHCPv6, DDNS, random cloud services, etc.).
Yet another reason why ipv6 will never happen
I think the Matter standard is going to cause IPv6 adoption to increase significantly in the coming years. People will demand it, without even knowing what Matter or IPv6 are. They just want to be able to turn their lights off from their phone without any extra hardware or software.
Matter is only local networking. Essentially everyone has IPv6 on their local network.
That said, I disagree with the parent statement that v6 isn't going to happen. It's already happening at a steady rate. From the Google stats it's steadily rolling out. Some countries are reaching 100% v6 deployment. I wouldn't be surprised if we eventually see some countries with high v6 deployments see services which are v6 only.
> Some countries are reaching 100% v6 deployment
World map:
https://www.arcep.fr/cartes-et-donnees/nos-cartes/ipv6/carte...
France is extremely close to 100%:
https://www.arcep.fr/fileadmin/reprise/observatoire/ipv6/Arc...
> some countries with high v6 deployments see services which are v6 only.
IIRC this has already long happened in some countries with smaller IPv4 pools, can't recall where.
They've had years to demand it, I don't see why it would suddenly change. Most users who buy "smart" appliances are not power users, don't set everything up themselves and just rely on manufacturer's servers anyway and don't need a direct connection - and are okay with it.
it already happened.
also I sure as hell dont remember my ipv4 address
I know your ipv4 address, 127.0.0.1. :)
There’s something to be said for human readable addresses. I’m a little nostalgic of how the .hack world was envisioned, where servers had address names like Hidden Forbidden Holy Ground.
If roughly 10 million words exist, then allowing any three words in order creates a space for 10^21 addresses… five words and you’re close to ipv6 address space, six words and there’s more combinations than ipv6 addresses.
Scrabble is 250-280k uk edition - wouldn't want to go too much beyond that I suspect. Where'd 10 mil come from?
Across 7000 languages in the world including non-latin alphabets…
I also know your IPv6 address, ::1
Even easier.
I don't remember my ipv4 addresses either, but I recognize them as mine when I see them.
I've never recognized an IPv6 address or prefix used in my networks.
You can set your ULA to something like "fddd:192:168::/48" and then on your vlan you prefix hint, say, "66". Now, any device on that vlan will be addressable by "fddd:192:168:66::$host". For example, your gateway ('router') for that vlan would be "fddd:192:168:66::1".
If you want to be really wonky you can script DHCPv6 to statically assign ULA IPv6 leases that match the IPv4, and expire them when the IPv4 lease expires, but like said upthread, addressing hosts via IPv6 is the wrong way to go about it. On your lan, you really want to be doing ".local" / ".lan" / ".home".
> addressing hosts via IPv6 is the wrong way to go about it. On your lan, you really want to be doing ".local" / ".lan" / ".home".
.local is fine as long as all the daemons work correctly, but AFAIK there's no way to have SLAAC and put hosts in "normal" internal DNS, so .lan/.home/.internal are probably out.
> On your lan, you really want to be doing ".local" / ".lan" / ".home".
The "official" is home.arpa according to RFC 8375 [1]:
[1]: https://datatracker.ietf.org/doc/html/rfc8375It may be the most officially-recommended for home use, but .internal is also officially endorsed for "private-use applications" (deciding the semantics of these is left as an exercise to the reader): https://en.wikipedia.org/wiki/.internal
That is a classical "design by committee" thing.
".home" and ".lan" along with a bunch of other historic tlds are on the reserved list and cannot be registered.
Call techy people pathologically lazy but no one is going to switch to typing ".home.arpa" or ".internal". They should have stuck with the original proposal of making ".home" official, instead of sticking ".arpa" behind it. That immediately doomed the RFC.
I do it by abusing the static slaac address. I have a set of wierd vms where they are cloned from a reference image, so no fixed config allowed. I should have probably just have used dhcp6 but I started by trying slaac and the static address were stable enough for my purposes so it stuck.
How does that work? I initially assumed you meant you just statically assigned machines to addresses, which I think would work courtesy of collision avoidance (and the massive address space), but I can't see how that would work for VMs. Are you just letting VMs pick an IP at random and then having them never change it, at which point you manually add them to DNS?
Pretty much. A given mac address assigned in the vm config maps directly to a static slaac address(the ones they recommend you not use) and those preknown slaac address are in dns, Like I said, I should probably use dhcp6 but it was a personal experiment in cloning a vm for a sandbox execution environment. and those slacc address were stable enough for that. every time it gets cloned to the same mac address it ended up with the same ip6 address. works for me, don't have to faf around with dhcp6, put it in dns. time for a drink.
But the point is that is the address you would put in dns if you also wanted to use slaac. Most of the time however you will just set a manual address. And this was with obsd, where when slaac is setup you get the slaac address and a temporary address. I don't really know what linux does. Might have to try now.
Clarification for others: with privacy extensions disabled, SLAAC'd IPv6 addresses are deterministically generated based on MAC addresses. There's also an inbetween (IPv6 are stable per network by hashing).
I run split-horizon DNS on mine. By far my preferred solution.
I don't even bother with split-horizon. My internal hosts are in public DNS under a subdomain.
it will naturally happen when you work with it long term, similar to how it was with v4.
I remember almost all of mine, including /24 and /16 blocks from several former employers!
Adoption has been flat under 50% for a year now.
A bit less than 50% of the whole Internet.
Wow. That's so amazingly unpopular. Why anyone bothers talking about something untold millions of people use every day is beyond me.
It’s not flat: https://www.google.com/intl/en/ipv6/statistics.html
You have to take into account seasonal trends. The summer is always higher, so yes, we’re currently below last summer, but we are above last April 1st, and this summer will be higher than last summer.
Also just looking at the history it always seems to have flat periods with sudden boosts which never go back down.
It looks flat to me since July of last year. Regardless, when you extrapolate that curve, when do you estimate hitting 100%?
> It looks flat to me since July of last year.
That was the entire point of my comment.
You can’t compare fall winter and spring to last July. You have to compare last July to this July, which hasn’t happened yet, but when it does, it will be higher than last July. Today (April) is higher than last April.
The reason for this is that more people are on mobile connections during the summer (kids home from school) so the summer (as well as the Christmas/new year week) are the highest points of the year.
The fact that it’s “flat” since last July, the high season, means it’s actually still increasing.
> Regardless, when you extrapolate that curve, when do you estimate hitting 100%?
Never? But what’s your point? IPv6 is a failure if it only replaces IPv4 for 99% of traffic?
I’m always bemused when I read comments like this. Regardless of whether you think IPv6 is good or bad, it is happening. IPv6 traffic to Google goes up by a few percentage points a year, steadily, and is at around 50% now.
Go 128-bit or go ::1
They should charge a small annual fee and let people reserve a custom word for a given IP. You could even have a small utility on your computer that automatically queries given names to "resolve" to IPs.
they should also charge a reasonably larger annual fee for reserving a short suffix and allow re-selling custom word prefixes..
To facilitate client to client communications, they should add a who-does-it-serve (short WHO DIS) system to get contact information of already reserved names.
And then if they don't like someone they should revoke their word. And they should make it so words with certain endings are delegated to other countries who can also revoke your word if they don't like you.
As long as there is a way to inappropriately link the security of my entire enterprise into it!
I tried it. Maybe it's easier to speak than hexadecimal is.
But I'm not sure that "How morally the enviable assistances categorize the insistent iodine beyond new time where new systems stalk" has the same memorable quality as "correct horse battery staple" does.
Reminds me a bit of S/KEY (RFC1760, RFC2289 and others around the 1990's).
Not because of the encryption element, but the part about representing a 64 bit integer as a six word sequence for usability.
(https://en.wikipedia.org/wiki/S/KEY#Usability).
Also used outside of that for quickly/easily recognising hash fingerprints.
(It's easier to recognise that your fingerprint is "GAFF WAIT SKID GIG SKY EYED" than "87FE C776 8B73 CCF9").
(It also slips some parity in there for good measure).
Being essentially impossible to memorize is one of the worst attributes of IPv6. I memorize and manually type IPv4 addresses all the time and it's super useful.
I've been memorizing and typing IPv4 addresses too, but I have enough devices on the network now that I can't remember nearly all the IPv4 addresses.
So then I need to use DNS. At which point it could be IPv6.
I have 56 host entries in my dnsmasq.conf.
It is notable that an IPv4 address expressed as a decimal number has up to 10 digits, the same as a phone number in many countries.
DNS, Avahi are super usefuler.
Good luck when you're trying to troubleshoot and DNS not working is one of the symptoms. 8.8.8.8 and 4.2.2.x are easy to remember.
So is 2620:fe::fe for Quad9 DNS
DNS should be auto configured and work with multiple redundancy these days.
If it breaks, so much that you cannot do a dig, you need to re think your network.
My DNS "server" is a router which can "add" static entries. Easy with static addresses, won't work with dynamic addresses.
What redundancy, multiple servers? Do you think everybody runs dedicated homelabs to access a raspberry pi.
> My DNS "server" is a router which can "add" static entries...won't work with dynamic addresses.
Sounds like a pretty poor setup, systems which could auto-add DHCP'd or discovered entries have been around for literally decades. You're choosing to live in that limitation.
> What redundancy, multiple servers?
Multicast name resolution is a thing. Hosts can send out queries and other devices can respond back. You don't need a centralized DNS server to have functional DNS.
Oh yes, that's really convenient for home users. "Install this thing on several computers and keep it in sync or you're not qualified to have a network"
Home users would ideally be served by things like mDNS and LLMNR, which should just work in the background. If I want to connect to the thermostat I should be able to just go to http://honeywell-thermostat and have it work. If I want to connect to the printer it should just be ipp://brother and I shouldn't even need to have a DNS server.
And if DNS fails, I have to use a serial console to get into my router and fix it, because I can't remember what address to type in ssh?
Your interface has a default gateway configured for it, doesn't it? Isn't that default gateway the router? NDP should show the local routers through router advertisements. There is also LLDP to help find such devices. LLMNR/mDNS provides DNS services even without a centralized nameserver (hence the whole "I shouldn't even need to have a DNS server"). So much out there other than just memorizing numbers. I've been working with IPv6 for nearly 20 years and I've never had an issue of "what was the IP address of the local router", because there's so many ways to find devices.
Even then nobody is stopping you from giving them memorable IP addresses. Giving your local router a link-local address of fe80::1 is perfectly valid. Or if you're needing larger networking than just link-local and have memorable addresses use ULAs and have the router on network one be fd00:1::1, the router on network two be fd00:2::1, the router on network three be fd00:3::1, etc. Is fe80::1 or fd00:1::1 really that much harder to memorize than 192.168.0.1 or 192.168.1.1 or 10.0.0.1, if you're really super gung-ho about memorizing numbers?
really home users who mess with DNS settings? Lot of people here are living in a bubble.
OK, so use the IPv6 endpoints? Write them down if you have to use them that much?
- 2001:4860:4860::8888
- 2001:4860:4860::8844
If you hate typing that much, computers may not be for you.
I really don't think 2001:4860:4860::8888 is as easy to remember as 8.8.8.8, no.
> If you hate typing that much, computers may not be for you.
Nobody said anything about typing?
What would have been your solution to needing more bits? More information is always going to be harder to remember.
Adding two extra bits to each octet, making each octet range from a still memorable 0-1023 rather than 0-255, would result in an addressing scheme 256x larger than all of IPv4 combined. The entire internet works fine even when IPv4 was nominally exhausted. NAT and CGNAT are not sins, they're not crimes, and there's no rational reason to be as disgusted with them as IPv6 fans are. Even then, IPv4 exhaustion wasn't really a true technical problem in the first place, it was an allocation problem. There are huge /8 blocks of public IPv4 space that remain almost entirely unused to this day.
The reason I'm an IPv4 advocate in the IPv4/IPv6 war is that the problem was "we're out of address", not "your thermostat should be natively routable from every single smartphone on the planet by default and inbound firewalls should become everyone's responsibility to configure for every device they own".
CGNAT is a feature, not a bug. Blending in with the crowd with a dynamic WAN IP is a helpful boost to privacy, even if not a one-stop solution. IPv6 giving everyone a globally unique, stable address by default is a regression in everyone's default privacy, and effectively a death sentence for the privacy of non-technical users who aren't capable of configuring privacy extensions. It's a wet dream for shady data brokers, intelligence agencies, organized crime, and script kiddies alike - all adversaries / attackers in threat modelling scenarios.
IPv6 adds configuration surface I don't want. Privacy extensions, temporary addresses, RA flags, NDP, DHCPv6 vs SLAAC — these are problems I don't have with IPv4. More features means more opportunities to footgun with misconfigurations, being forced to waste my time learning and understanding the nuances of each (in again, what amounts to system I want nothing to do with).
"Reaching your own stuff" is already a solved problem, too. Tailscale/Headscale gives you authenticated, encrypted, NAT-traversing connectivity. It's better than being globally routable. It's also opt-in for anyone who wants it, and not forced on anyone, unlike the IPv6 transition.
Saying that IPv4 is ok because we have NAT and CGNAT is like saying that spam is not a problem because we have spam filters everywhere.
I don't have your problems with ipv6, and I'm actively using it.
I don't have to rely on extra commercial entities to be able to reach my network.
I did have a problem with hosting my own shit because my ISP by default does cgnat. That cost me an hour of my life to convince a party to give what used to be normal, end to end connectivity.
The tailscale client and the headscale server are both open source, you don't need to rely on commercial entities.
yes you do, the control plane is closed. Only reverse engineered by the headscale project. The control plane is necessary for the peers to find each other. If you need to rely on such a crucial part being reverse-engineered, than yes, I think it's fair to say you are ultimately relying on commercial entities.
Headscale is open source and it already works. You don't need to rely on anyone to use it, or even to improve it.
Wouldn't easy and accessible self-hosting be a major privacy win if that's your primary concern? Sounds much more private to run a Minecraft and Mumble server on an old laptop in a friend group than paying a commercial entity like a hosting provider to know about it and have a back door.
Easy and accessible self hosting isn't the primary concern.
It's much more private and secure to run that Minecraft or Mumble server on an encrypted overlay network like via headscale + tailscale rather than exposing both services directly to the entire planet.
But again, the primary concern was only ever address space.
What I tried to express was privacy being the primary concern. The easy and accessible self-hosting on old hardware would be the uses of a home network beyond superficialities like consumption and commerce. Privacy wise headscale as a solution is still not quite there, because it either necessitates an additional third party to host the headscale server and know about all my friends, or jank like dynDNS.
The additional security gained by getting everyone involved to set up and configure separate VPNs for different community utilities is not worth it.
I disagree wrt NAT. It creates huge problems for many p2p applications.
I wouldn't call port forwarding "huge problems". It's only one minor router setting and if you don't want to deal with it, there's the abomination called upnp.
> I wouldn't call port forwarding "huge problems".
Port forwarding has massive problems if you're running applications expecting certain ports and need multiple hosts to have public access to those ports.
>"Reaching your own stuff" is already a solved problem, too. Tailscale/Headscale
IPv6 predates those by decades.
> NAT and CGNAT are not sins
Highly disagree. Middleboxes are a huge problem on global scale and have frozen any innovation below application layer. TCP and UDP even that they are on software not hardware layer cannot be updated or changed, see MPTCP efforts or QUIC giving up and building on top of UDP.
If this is so much privacy problem, IPv6 is there for many years reaching 50%+ deployments in some countries, I bet there should be concrete examples of such breaches and papers written.
> Reaching your own stuff is already a solved problem, too. Tailscale/Headscale
No address to receive communication - no problem install an app that would proxy it through someone who has the address. Tailscale/Headscale is great, using it daily, but they are not solution to the huge already build global network created to connect devices not connecting devices because lack of digits. Global is key here.
It reminds me of what3words, using three words to describe any location on earth. I really hoped that could catch on.
Kind of like what3words, except what3words uses three words which you stand a chance of remembering, whereas this produced, for an address similar to mine, "Miniature nerves eulogize gaily inside erect lion yet able stables hiss the conclusive consultation."
what3words Is terrible and search and rescue teams are actively against people using it.
Care to elaborate why? I'm just curious since I didn't know (1) there was actually any kind of serious usage, or (2) there was pushback from rescue team...
W3W is very aggressive about protecting their IP, they don't want it to be a standard anyone can use like lat/long.
They advertise it as being useful for search/rescue as you can provide a precise location over an unclear voice channel. They conveniently ignore that speaking numbers is clearer than speaking random words.
I'm sure there's more I'm unaware of.
One reason is that there are a lot of similar words in the dictionary. It is easy to mishear the wrong location especially when they are close together. Some of the words are long and complicated. Another is that they are random which means can't navigate from the codes.
They are missing feature of some codes that can have variable length for variable precision.
> what3words Is terrible
Yes
> search and rescue teams are actively against people using it.
Sadly no
https://media.neas.nhs.uk/news/3-words-can-save-lives
seems like they’re quite against it going by what i found on reddit. https://www.reddit.com/r/searchandrescue/comments/1d63v94/wh...
Something that I think was probably once obvious to me but I rediscovered recently is just how intensely wired for song the brain is. If you want to memorize anything, doing it as a song makes it far easier.
I’d really love to see things like this generate little jingles along with the sentence. :)
There actually was an attempt on HN a little while back to use GenAI to convert facts, flashcards, lists, etc. into automated melodic mnemonics. The biggest issue in that particular case was that it was also generating the motif from scratch.
At least for me, part of the reason I can still sing the countries of the world is because the original Animaniacs song was set to a tune that was already familiar: “Jarabe Tapatío” (aka the Mexican Hat Dance).
I memorized that (and several other) Animaniacs songs without being familiar with the melody. Even Tom Lehrer’s The Elements reached me before Pirates of Penzance did. I think the melody just needs to be simple, then it'll become ”familiar” quickly.
However, for the use-case at hand (remembering IPv6 addresses) I don't think I'd use that. I'd just write them down somewhere, like, uh, perhaps, oh I know: the hosts file.
“Catchiness” is probably more important than anything, hence the concept of the earworm aka stuck song syndrome. Even SOTA GenAI like Suno/Udio fall pretty short of generating genuinely engaging melodies.
As anecdotal evidence, I could never remember all the names of the london tube stations as prose, but after listening to Jay Foreman's Every Tube Station song a few times, I cant help but sing it every time I hear a tube station name, and can almost flawlessly list them all off
https://youtu.be/8jPyg2pK11M
It's huge when you consider all the data humans have stored and transferred orally over the millennia.
Music, meter, and rhyme are all (among other things) algorithms for indexing and error-correction, tools very suitable to the squishy hardware.
The new times take now beneath the new time while new times take the new year.
Or more concisely, localhost.
Commented with the same without reading through the comments first; now deleted.
That is a lot of words for ::1.
Edit: there was another one below, haha :)
I’m afraid to ask, but why, and who, tries to or wants/needs to remember IPv6 addresses?
What is the use-case for this? I’m trying to think of an IPv6 address I would need to remember, and then when I’d have access to this site without having access to a text file where I could have noted the address down. I’m coming up empty.
it may more likely have something to do with the date of April Fools :) seems it got more than a few folks here haha
It'd be nice if simpler addresses gave simpler sentences. fe80::1 translates to "Uninhibited times take now inside new time yet new times take the new year."
I would have enjoyed a blog entry detailing how this works, regardless of its practical utility.
"The amazing champions inspire boldly like brilliant genius and incredible legends admire splendid talent."
Hard to forget a sentence like that!
I don't understand how the mapping works. An address has 8 parts and produces 16 words, so each part consists of 2 words. If we take the example 2a02, that gets encoded to "how atop", but I don't see how that text helps me that "how atop" means 2a02? Am I suppose to memorize both? How does that help?
You are not supposed worry about the mapping. You trust the website to help decode it. You just remember the sentence. It's a little like what3words for coordinates.
The rationale being you are more likely to remember grammatical cogent sentence, than a random string of alphanumeric characters. Although I will agree that the generated sentences don't seem easy to remember. So I doubt it's utility.
The new times take now beneath the new time while new times take the new time.
So just imagine if there was a service that could translate any words you wanted into the IP address instead of relying on some website to generate jibberish. Wouldn't that be cool to use instead? Some kind of name system? Based around domains of authority?
http://borg.uu3.net/~borg/?ipv6
Now, if only those people who designed IPv6 were smarter.. Hex aint that bad, LONG hex addresses are pain to use.
Now, lets say you have LAN like this [::1:0:0/56]. So, ::1:0:24 is easy to remember right? Managable? right?.. Also, bonus for :: shortening is, you immediatly know what are you dealing with, ::1 is loopback, ::1:1 is LL, ::1:0:1 is LAN.. everything else is Internet.
The truth is, IPv6 is really 64bit, the other 64bit part is just randomish node address...
> The truth is, IPv6 is really 64bit, the other 64bit part is just randomish node address...
So anyway it gives 128bits in total, 64 for network and 64 for node.
But I wish there was a better way to write just the local node part and global part being taken automatically.
This encoding is so long, that I'm more likely to remember the raw address. :D
And I don't think I ever typed manually any IPv6 address other than `::1`.
No need to type `::1` anymore, you can instead just type `The new times take now beneath the new time while new times take the new year.`
OK. That's much easier. :D
You can make unique local address subnets with simple addresses.
I use a simple one like `fd10::1/128` and `fd10::2/128` and so on
Technically speaking RFC 4193 says you should use random bits. But I don't care.
The first (of two) examples encodes to:
> How now the smart flies take the new time beyond new time where new times come.
..Nice idea, but it may need some more thought. (Even more so as 2001:db8::1 is much easier to remember than that!) (I wrote that parenthetical from memory on edit, vs. had to copy-paste the sentence when it was my intention to comment on it within seconds.)
I'm old. I can't remember breakfast.
why is there even a need to remember IPv6 address in the first place?
Not too sure of the utility of this. It's not an easy sentence to remember, because while grammatical, it's nonsense—it would take some effort. So if I'm trying to memorize a static IP, setting up a DNS name is likely to be easier. And also if I'm going to be using this to memorize IPs I'd like the algorithm to be open source.
All that being said, I think it's a neat idea and a cool tool!
love to get an api for this.
Ah yes, because "How now the smart flies take the new time beyond new time where new times come." is so much easier to remember than "2001:db8::1".
We kind of had the same idea for ECDSA public keys (an imagined solution to zokos triangle -- human readable and decentralized) as well as private keys (BIP39 brain wallets). Honestly it still falls short of truly name-based though.
Just proves that 16 bytes was too much, and we should have just gone 8 bytes.
I'm fine with 16 but they should have only used the bytes as they were needed, at least for 5 and 6 byte addresses, so those who desire short addresses could buy them.
At least on a LAN, you can set up like fe80::3 . I think. Now I'm not sure if I got that right.
fe80:: is for link local. You'd want to use something starting fc00:: or fd00::
In your typical home environment, just set your ULA to fd00::12 instead of 192.168.0.12, or fd00:16:34 instead of 192.168.16.34
Yes you'll run into issues if you were to later want to merge your private IPs with someone else, and you should use fd12:3456:7890::12 instead, remembering those extra 10 digits, but its not a problem at home, and no more of a problem with business mergers than ipv4 clashes anyway.
My vrrp address for my dns server at home is 2001:8b0:abcd::53
It's about as easy to remember as 81.187.123.45//192.168.0.53
Almost all ipv6 addresses I encounter start with 2001, so I just need to remember my home prefix is 8b0:abcd, which is about the same length as my home public IP of 81.187.123.45
::53 means subnet zero host 53, which is easier to remember than which rfc1918 range I used (and basically is the equivalent of remembering the 2001:: prefix)
If I have an internal server I'd have on 192.168.4.12 I could address it with 2001:8b0:abcd:4::12 just as easily, with the "4.12" translating to "4::12", and the "81.187.123.45>192.168.x.y" translating to "2001:8b0:abcd:x::y"
Just because slacc gives you an extra 64 bits of randomness doesn't mean you need to use them.
Mine comes with a swear!
[…] thaw the new case beyond pure mass where flagrant toys fucken.
ipv6 is for faceless hordes of cellphones, which could just as easily be NAT
despite being an ipv6 skeptic, i’ve been thinking to try using ipv6 for our new company network, but make the addresses purely readable
If you're assigning addresses, you can make the addresses in a ULA as short as you want. You're supposed to use a random 40 bit network id but if you can accept that you may need to renumber at some point there is no reason you can't use fd12:b:a:d::beef or whatever.
Yes I'm going to generate random numbers until the number I generate is fd77::
There's another way to make addresses purely readable that's been around longer than NAT: DNS.
> is for faceless hordes of cellphones
How could we determine which device on mobile network is a faceless cellphone and which is a proper device needing real sweet Internet connection? And won't that make things more complicated than just v6 deployment?
Can argue that NAT, which interrupt layers ment for end device do basically the same as popular user hostinle unchangable mobile OSes, but I don't think latter is good either.