IPv6 Real World Maturity (was re: How long is your rack?)

Owen DeLong owen at delong.com
Mon Aug 15 22:07:03 UTC 2011


On Aug 15, 2011, at 2:14 PM, Tim Wilde wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On 8/15/2011 2:24 AM, Owen DeLong wrote:
>> What does it say that the same thing happens in IPv4?
>> 
>> I really don't see a significant difference in that regard.
> 
> I will admit to not having run the numbers and trying to compare IPv4
> protocol-specific discussion threads vs. IPv6, but it certainly "feels"
> like there are more.  My feeling is also that the IPv6 discussions are
> much more fundamental, in that they're discussing basic deployment
> strategies, etc.  But it could all be selection bias because it's
> prominent in the collective mindset, I'll grant you that.
> 
I was talking about quality, you're talking quantity. Sure, there are more
IPv6 protocol discussions, it's a newer protocol, there are more people
left that haven't had all of the same old discussions, haven't gained some
experience and come back to the same old discussions with new perspectives,
etc.

However, the quality of the IPv4 same old discussions vs. the IPv6 same
old discussions is roughly the same. It's all about problems or perceived
problems that we knew about from very early in the protocol's design
life and somehow the protocol works well enough for lots of people to use
it in spite of these (seeming from the discussions) overwhelming flaws.

As an example, look at how often the NAT != Security / Yes it does.
argument still comes up in spite of the fact that it's been pretty clearly
established that NAT is actually neutral at best and usually detrimental
to security, while it does offer some small privacy advantages.

Lately, I'll admit, that argument comes up most often as part of a "but
what do we do in IPv6 without NAT? All my windows boxen will be
exposed naked to the world?" discussion, but, I'd say that's still an IPv4
discussion, not an IPv6 discussion. Without the damage done to IPv4
by NAT, we wouldn't have people who grew up not understanding
how networks are supposed to work and unaware that stateful
firewalls can work just as well without NAT as with.

>> Yes, IPv6 is currently a little less fully baked than IPv4. IPv4 is
>> 20 years older than IPv6, so I say that's to be somewhat expected.
> 
> Point taken.  Anyone have time to try to do a long-term comparative
> study of discussions on deployment strategies and things like NAT, DHCP,
> etc, for IPv4 vs. IPv6, factoring in the differing levels of overall
> Internet adoption at the time of IPv4 adoption vs. IPv6, etc?  If so, I
> have a few other tasks I'd love to have you do... :)
> 

I don't think that's a relevant question. At the time of IPv4 adoption,
the internet didn't have WWW or HTTP or much in the way of end
users. IPv4 was adopted when SMTP and FTP were the primary
applications with the occasional telnet. I think at that time, there was
almost as much ping and trace route traffic as anything else (ok, not
literally, but you get the idea).

However, given that 30 years later, the quality of the IPv4 same old
discussions is on par with the quality of the IPv6 same old discussions
and IPv6 only wins on quantity at the moment because it's new, I'm
not sure anyone really needs a study to confirm that. However, if there's
a researcher out there with too much time on their hands, go for it.

> As others have said, I guess what it really shows is that nothing ever
> really changes, and no one (protocol designers, IETF folks, operators,
> router vendors, etc) is perfect, despite our best efforts to be. :)
> 
Yep.

Owen

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 2105 bytes
Desc: not available
URL: <http://mailman.nanog.org/pipermail/nanog/attachments/20110815/e0d63c41/attachment.bin>


More information about the NANOG mailing list