IPv4 wealth redistributed IANA has published its six-monthly review of IPv4 address recoveries and reallocations. The advice reveals the following data on who got what, illustrated in the table below. The world's IP-address-handing-out authorities run their own queues with their own rules for dishing out addresses, so the fact some addresses have been …
"For the rest of us, getting on with IPv6 remains as sensible as it has for years."
And when will the Reg be IPv6? For even a test? A beta? Anything?
If you're going to be putting snarky comments on IPv6 or SSL/TLS articles still, make sure you aren't one of the exact people that that snark is aimed at.
Well, I had v6 - I turned it off.
When you have retards like Google and Microsoft (probably the same person in the Gmail team getting a job in Azure after that) using v6 in a small network is a lose-lose proposition.
The RFC-specified email retry semantics are - you try the lowest MX, if that does not work, you try the next higher and so on. It says _NOTHING_ about v4 or v6.
The retarded monkey in the Gmail team who coded their retry sequence _LEAVES_ it on v6 and does not try v4 ever again. Once v6 always v6.
In violation of all RFCs. So if you have two links - one with v6 the other one without you lose mail delivery retry capability for anything coming from the Chocolate Factory and will occasionally lose mail. Outlook 365 moved to a similar bogus retry sequence at some point too.
As a result even Comcast which is one of the biggest proponents of v6 amidst SPs has _REMOVED_ their v6 MXes and has switched back to v4 only mail delivery.
So on the balance of things - a network transport which is not usable for the most basic Internet protocol? You call that sensible? Give me a break.
The IPv4 Internet is where everything you want to talk to is. The parallel IPv6 Internet has almost nothing, bar Google and Facebook.
So which should you do?
1. connect to the IPv4 Internet
2. connect to the IPv6 Internet, and then find you have to connect to the IPv4 Internet as well
In what way is option 2 more "sensible" than option 1? And how does it reduce your requirement for IPv4 addresses?
What is lacking is any sort of interoperability between IPv4 and IPv6.
@AC
Obviously you can still connect to the IPv4 internet - after all its going to be around for a while - but it can't keep growing due to the lack of addresses and finer slicing of the smaller free pools is just increasing the size of the global routing tables, which has a knock on effect on the routers that need to process that data. The routers have finite hardware resources to hold and process this data.
Given that you can't cram any more information into 32 bits, then more bits are required to solve the problem, hence the incompatibility between the two versions. This is why its always been called a period of co-existence and why your OS has two IP stacks.
There are some interoperability techniques, that work with different levels of success and different limitations, but don't confuse interoperability and co-existence with actually migrating. The purpose of the dual-stack approach is to allow a painless migration from the older standard to the newer one.
Those who fail to migrate will be the ones left out in the cold once the V4 pools dry up, at which point the pain will really start. New Internet resources won't be able to come up without adopting a V6 only address and without users enabling IPv6, users won't be able to access those sites, so both groups need to do their bit. This is why its called co-existence and parallel running.
As to your point about "IPv4 internet is where everything you want to talk to is", that may have been correct in years gone by, but only because V4 is the incumbent IP version running everything to date, but that's not the same as the "V6 Internet being empty". There are large swathes of IPv6 enabled sites out there.
Look at RIPE's graph of V6 utilisation Here
Also look at Akamai's report of what they see Here
The only thing that is changing is that the time before we hit the wall is decreasing.
I did some random tests just now and IPv4 appears dominant because most web sites are using edge caches. Those don't consume a unique IP address for each site so they're probably in no hurry to upgrade. The Reg is using double-agent CloudFlare.
The biggest problem I see is that most routers are complete garbage. It says "IPv6" on the box but it doesn't work. The best I've seen so far is a functional switch that enables and disables incoming WAN connections. Forget about attempting to use firewall rules. I mean, most small business routers got support for multiple WAN IPv4 addresses working just a couple of years ago. Give Cisco/Linksys/D-Link/Netgear another 20 years for IPv6.
IPV6 is just not human friendly and that's it's biggest problem. People can rattle off IP's (well some can) whereas I would never be able to remember a v6 address. Not only that but there's going to be some surprises for people who misconfigure their routers and give their internal kit/machines publicly routable IP's without realising. Redistributing the IP4 wealth should be exactly that, doing a proper full audit on usage of existing blocks and releasing chunks that are never used. IPv4 is not going away for a very long time, it's user friendly and simple. Let it be used for IoT stuff that doesn't need human involvement.
I thought that these computers were supposed to be looked up by name! What is the point of DNS when everybody tries to refer to a number? (I run into this at work all the time. "What's the server?" "It's blah.blah.com." "But what's the address?" "We have DNS. Use it." So they do a query on the command line, and then use the IP address...)
Agree 100% on the DNS point and worse than that, hard coded IP addresses prevent any form of sensible resilience and disaster recovery processes from working properly, so hard coded addresses are a major operational risk.
The root of this problem is that people don't update DNS when they make changes to the Internal infrastructure, so its a home-grown issue within IT teams. Conversely who types in an IP address to get to a public site like Google's - nobody, since DNS works perfectly when its zone data is up to date (there's a hint here for people using addresses).
I've also found it odd that remembering random IP addresses is support norm, but if you ask programmers what all the hex codes are for the CPU instruction set, they will give you a blank stare and will state that's the compiler / assembler's job to remember.
Which implies your internal DNS is not resilient or is incorrectly set up, which was my previous point.
DNS is resilient and reliable if deployed correctly, which we know from using the Internet for many years.
Spend the time fixing any internal DNS mis-configuration issues, rather than trying to remember and hard coding static IP addresses. This is not an IPv6 issue, its a bad practice affecting how people bung in DNS, then try and work around that bad practice.
And when presented with one of those misconfigured DNS installations, you never have to look at, type in, or mentally compare an IP address?
This reminds me of calling tech support because your computer won't boot, and you're told to fill out a trouble ticket on their website. The high-level solution is best until it isn't there.
Human-readability of addresses isn't the most important issue (by a long shot), but it sure is annoying.
The IP6 wasn't designed for the real world, easy limiting / blocking traffic, privacy or security. Just to be big and fancy.
It's been fudged a bit regarding original MAC scheme, but it's flawed in so many ways. They need to think about it again, it's a stupid system given long term need of IP4 (many gadgets can't be updated or never will be). No important Website can turn off IP4 till almost no-one is using IP4, there are better ways of extending address space.
The world is a lot more than Servers and Corporate IT.
Even if my ISP supported it, many gadgets don't and it's a horror for anyone other than Network specialist to setup, I mean more so than IP firewall/NAT/Security Router & home WiFi, which is bad enough.
Well, Google would love that IoT (or more sensible WiFi gadgets that don't need Internet Cloud / 3rd Party) only uses IP6 as home users would NEVER figure how to block unwanted exploits from Security forces, criminals and Cloudy Exploitive Commercial companies.
I enabled V6, after my ISP (Zen) opened a trial. Yes, it's 2016, and it's just a trial. However it took nearly a month for them to set up my reverse DNS entry, and as most websites are still IPv4 anyway I found little benefit.
I have a backup network solution, via EE 4G, which offers no IPv6 at all. So rather than have a dog's breakfast of stuff on my LAN, I reverted to IPv4.
This post has been deleted by its author
Our isp doesn't offer ipv6 at it's data centre let alone on it's broadband network.
A number of my data centre/cloud providers don't support ipv6 in their control panels though most do not have ipv6 support. Even our domain registrar (domainmonster) don't support ipv6 (you can't for instance add an ipv6 glue record) although they can do it manually for you.
Also I've just checked again and see google still don't have ipv6 addresses for their name servers.
All in all ipv6 doesn't seem to have made much progress since ipv6 day in 2011.
@AC
Google's IPv6 addresses are available and have been since at least 2011. See Google DNS, which states
The Google Public DNS IP addresses (IPv4) are as follows:
8.8.8.8
8.8.4.4
The Google Public DNS IPv6 addresses are as follows:
2001:4860:4860::8888
2001:4860:4860::8844
You might note the similarity between the addresses and that they are only 4 chunks long, like their IPv4 counterparts, so fairly easy to remember..
As to your other points, change - its simple - change your suppliers to those that do support IPv6, there are plenty out there.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
