On killing IPv6 transition mechanisms

[Ted Mittelstaedt]

Gert Doering wrote:
> Hi,
> 
> On Thu, Mar 18, 2010 at 11:02:43AM +0000, Benedikt Stockebrand wrote:
>>>> ok, just to rephrase my original question: If you connect to the
>>>> Internet via an IPv4-only ISP (rather than one that offers native
>>>> IPv6) and then alternatively use native IPv4 and 6to4 to access a site
>>>> serving both, do you still notice a difference in latency?
>>> I'm pretty sure that I will.  That was the initial starter for this 
>>> thread - 6to4 is usually worse than native IPv4, and is causing real
>>> problems.
>>>
>>> Which is why it is fortunate that all the (recent?) Windows versions do 
>>> the right thing, and prefer native IPv4 before 6to4 and Teredo - so 
>>> for connecting to a dual-stacked hosts, 6to4 latency is irrelevant.
>> we're still talking different things here: If you are on the
>> *customer* side and connecting to a dual-stacked host when you are
>> using 6to4 yourself, then you do see the difference.
> 
> If you connect to a dual-stacked host, and your operating system is 
> behaving as Windows does (prefer native IPv6, but if that is not available,
> prefer IPv4 before 6to4 and Teredo), your client will not use 6to4.
> 
> So where should the difference come from?
> 
> (If you connect to an IPv6-*only* host, of course it will have to use
> 6to4, and that might be lots slower than the IPv4 path, but then, there
> for an IPv6-only host there is no v4 path to start with).
> 
> We are talking about the same thing, but you are ignoring the default
> precedence tables in current Windows versions :-)
> 
> 
>>>> Because its "kewl" and/or they want to see if it actually works.
>>> We know that IPv6 works.  Since about 10 years. 
>> Well, six or seven years ago I first had to tell people how to
>> recompile a kernel on Linux, download some extra software for XP or
> [..]
> 
> Yes, I agree.  The client and application side has become much less
> painful, and much more automatic in regard to IPv6.
> 
> And of course, things still break - which why it is so important to
> start now (well, "to have started 5 years ago") to be ready in time.
> 
> [..]
>> What we need are those enthusiasts spending 500Euro for a new graphics
>> card, tinker with absolutely everything they find and are generally a
>> major pain in the rear side for everybody else---but are still asked
>> for computer-related by their friends and family.
>>
>> We don't need them for testing but for marketing IPv6.
> 
> "Roll out IPv6, make 'the ping' slower for IPv4 than for IPv6, that
> will get their attention" :-)
> 
> Latency *does* matter!
> 
> [..]
>>> By replacing them with much larger problems, like "all of the absolute
>>> URLs that your sites might use for links will no longer work" - ask the
>>> heise people what fun they had with www.six.heise.de - they had to 
>>> setup a special proxy that rewrites the HTML pages in-flight to get
>>> the URLs fixed.
>> Agreed, when we're talking web servers and URIs that's not the very
>> best move.  But I was actually thinking more along the lines of DNS,
>> NTP, CIFS/SMB, NFS, SMTP, POP3/IMAP and such.
> 
> I've run my machines dual-stacked with a single DNS name for dunno
> how many years, and all these services just work in a mixed v4/v6
> environment - older clients generally use v4-only, newer clients try
> both and fall back to v4 if v6 doesn't work.
> 
> Of course there are broken clients, and broken servers, but if you
> do things like "ntp.ipv6.my.domain" and then don't advertise it, you're
> just hiding the problem - it won't go away, just by pretending to the
> broken client that IPv6 doesn't exist.
> 
> [ Alternatives to IPv6 ]
> 
>>> Well, fortunately, multiple layers of NAT are still the exception -
>>> so the deployer of the NAT is still in control regarding port
>>> forwardings and so on.
>> As far as I can tell that largely depends on which part of the world
>> you're living in.  
>>
>> It's a shame you haven't been in Potsdam two years ago, there was a
>> Korean colleague there who could tell you more about the problems
>> running large NAT gateways at the ISP side than you wanted to hear
>> about.
> 
> I've seen the newly-presented Cisco NAT Services card, and have a vague
> idea what the list price might be for something that holds 20 million
> nat table entries...
> 
> 
>> And we're facing pretty much the same situation with 3G/UMTS even
>> here.
> 
> Yes.  But if you remember last year's IPv6 conference, at least Vodafone
> has seen the light regarding UMTS and packet clients.
> 
> 
> [..]
>> Still, my point was on IPv6 deployment in an enterprise.  Just
>> recently I got a training request from a customer with such an
>> enterprise.  Just to give an idea of the numbers: 3 data centers, 150
>> locations, 40 000 employees.  I don't have the number of desktop and
>> notebook PCs, but you may well assume in that case that just about
>> every employee has at least one computer.  In another case I had
>> participants in a training who were dealing with 8000 locations.
> 
> These enterprises are the ones that will be in for a nasty surprise
> when they roll out Win7 or Server2008R2, and all of a sudden they have
> uncontrolled IPv6 all over the place.  So it's quite important that they
> understand what is coming up, and get prepared.
> 

Since those orgs use prebuilt images when rolling that stuff out, they
always have the ability to disable IPv6 in the image then roll it out
and that is likely what they will be doing.

It's really going to be the midsize orgs not enterprise who will be most
affected because they tend to not do forklift rollouts of machines 
across the enterprise, but instead move machines around between people
and buy the machines one at a time as they need them and do not use
prebuilt images, they just use the load that comes on the system.  They 
are bringing win7 into the network now and mixing it with XP and some 
small amount of Vista.

Ted