IPV6 is a failure – stop wasting everyones time.
Today I can't reach pypi.python.org. Why? Well because of this:
warnerc01:~ cwarner$ telnet pypi.python.org 80Trying 2001:888:2000:d::a3...telnet: connect to address 2001:888:2000:d::a3: Host is downTrying 82.94.164.163...Connected to ximinez.python.org.Escape character is '^]'.
What is this horse shit you maybe asking? Well the 2001:888:2000:d::a3 is an ipv6 address and it's looking kind of like, it's unavailable. However after trying the ipv6 address we fall back to the ipv4 address and it WORKS. Why is this a problem? Well telnet is great because it falls back but other programs aren't as capable. Which means they don't fallback, which means I believe pypi.python.org to be completely unavailable. It's not though because we see a connect fine to 82.94.164.163 which is an IPV4 address. Why should we have to fallback anyway? None of this makes any sense.
Daniel J Bernstein, says it best:
It gets worse. The IPv6 designers don't have a transition plan. They've taken some helpful steps, but they typically declare success (``IPv6 support'') when the real problem---making public IPv6 addresses work just as well as public IPv4 addresses---still hasn't been solved.
You can read more here at "The IPV6 mess" which describes the problem in a rational, coherent and logical manner as to why IPV6 is a failure and wasting everyones time.
[1]: http://cr.yp.to/djbdns/ipv6mess.html
December 31st, 2010 - 18:10
I like stumbling upon comments like this. They can’t simply say no to IPv6 since IPv4 has stretched the internet about as far as it can go. Within the next few months we should see the last few million IPv4 addresses assigned and then there can’t be anymore. Yes ISPs have kind of failed to get IPv6 rolled out as quickly as they should and it could very well be ugly when the end comes but you can’t just say no to moving to IPv6, it simply isn’t logical.
January 10th, 2011 - 12:29
Yes, we can absolutely say no to moving to IPv6 and instead come up with the next logical and functional progression of the IP protocol. No one is saying that it doesn’t need to be IPv6. However there are architectural and progression problems with IPv6 that simply have not been addressed and it seems like they will not be addressed. The problem is to simply stop wasting everyones time and working towards a functional and usable plan based on marketing etc.
February 13th, 2011 - 17:27
We can come up with the next logical progression? Come up with? Now? IANA is OUT of IPv4 addresses. That’s it. We’re in the endgame now. IPv6 is supported by the current state-of-the-art routers, OSes, applications, and about 30% of ISPs, not to mention the large number of IT professionals who have now been trained in the subject, the years of testing that were done on 6bone, the rollout of DNS AAAA records and everything, all the time and effort spent on advocating the transition, convincing China to wire its universities for IPv6, etc. A whole new internet has been deployed, across hundreds of countries, at enormous expense. Google and CNN and dozens of other websites have made their sites IPv6-accessible, and dozens more (Facebook, etc.) have pledged to by Q4 of this year. Millions of pieces of new hardware and software have been deployed. It took more than 10 years, but we finally did all that stuff and now if you look at the graphs, the number of IPv6 hosts is finally increasing exponentially (starting in 2007). All the pieces are finally in place and things are finally happening after a ridiculous amount of pain and suffering. And NOW you want to throw out the spec and start over? And you think this unspecified, utopian vaporware of yours can be deployed on hundreds of millions of disparate systems in time for IPv4 RIR exhaustion in August-ish?
Get real. There are, at this point, exactly two choices that don’t involve sorcery:
1. Carrier-grade NAT and the horrific dawn of a two-tiered internet without P2P or home servers.
2. IPv6.
Yes, the design and planning of IPv6 was an unmitigated disaster. I, too, was one of those screaming at the monitor when I first read those RFCs way back when. But what’s done is done. 64 bits is a lot of space; we’ll have more efficient routing and the addresses will flow like wine. No more dynamic IPs, either. Of the two options, it is obvious which one is correct. We can whine about the past, or we can grow up, accept it, and worry about the future.
As for the future, how exactly would abandoning IPv6 advocacy help us? Can simply giving up render a world mired in NAT more bearable in some way that I’m failing to see?
February 13th, 2011 - 18:49
I’m not sure if you read my last comment or you just felt the need to rant for whatever reason. That aside, can you at this time explain the exponential increase of IPv6 acceptance? For the last two years there has been an overall increase of approximately 0.1% and that is in actual deployment of IPV6. So if it takes approximately 10 years to get about 0.3% overall deployed addressing what makes you think there is going to be some ground swell mass movement to ipv6 just because there are no more ipv4 addresses?.
Next can you feel free to explain this flow like wine statement. The no more dynamics ip statement? All of this reads like a truly uninformed pontificator because you read somewhere that ipv6 was good.
What about actual deployment?
What about the incompatibility issues?
What about the massive amount of work involved that still hasn’t been done?
What about the transition process for machines, routers and others that have been working fine for two decades?
I’m sorry, but you’ve addressed none of these issues except to rant about more logical spacing, wine flowing and years of testing of the ipv6 protocol which is simply not the issue. IT professionals trained on IPV6? Surely you jest.