A quick introduction and thanks!

[NewtonNet]

As a newbie to the service I thought I'd break the ice with a quick introduction.

It's great what you've created here HE, not only your tunneling service but just as importantly your certificate programme which serves as a great target to aim for in a structured way. As I'm sure we've all felt, when you first exercise the initiative to experiment with IPv6 there is a certain level of excitement to be had from getting it up-and-running (much to the bemusement of non-geeks I'm sure!) but the novelty soon wears off and you can't help but think 'okay; now what?'. Of course, the options are endless but it's nice to have something to set you off in the right direction.

So thanks again, and hopefully as my own knowledge increases I'll be able to give at least something back in these forums!  :)

Mathew

[jimb]

I consider myself a "geek" but was still excited about learning about IPv6, just 'cause it's something new.

And yeah, it's somewhat anti-climactic, but that's really a good thing.  Much about IPv6 should be familiar from IPv4.  IPV6 is really an evolution of IPv4, not a revolution.  It builds on the successful "formula" of IPv4 while solving the main problems of IPv4 (the biggest, obviously address exhaustion).

But in terms of everyday internet use, one rarely notices the difference whether your session is using IPv6 or IPv4, which is also a good thing.  I only know while web browsing for instance, because of a plugin I have for FF which shows the address I'm going to down in the corner.  And by watching my networking monitoring meters if I happen to glance at 'em.

Where things get really interested and somewhat complicated is in the transition technologies.  If this sort of thing interests you, you may want to read up on 6in4 (which you already are using), 6to4, Teredo, TRT, NAT64/DNS64, and Dual-Stack Lite (DS-Lite), the first three being ways to access the IPv6 internet from an IPv4 island, and the last three being ways of doing the opposite (which will be an increasingly common situation for end users as IPv4 addresses are exhausted).

[NewtonNet]

Wise words Jimb, and I think you're dead right - I shouldn't be noticing when I'm using IPv6 any more than anyone else notices they're using IPv4. Of course, like you, I also have that plugin (ShowIP?), so I do happen to notice and I must say I find it somewhat depressing (or worrying at least) how few IPv6-enabled sites there are out there, particular those that probably ought to be living and breathing the stuff (ISPs and the like).

Thanks for the pointers about what areas to look at studying next. I think that's partly my problem at the moment - IPv6 seems so vast[1] when you include the various transition methods, support functions, etc and to make matters worse many of them seem (to my untrained eye) to be competing with each other or, worse still, be obsolete ideas/concepts that are no longer in use!

Cheers,

Mathew

[1] I think this fact may well become one of the wakeup calls for the 'late adopters'... Indeed I've had colleagues at work ask 'which RFC [singular] covers IPv6?'...

[broquea]

Re ShowIP, sadly this only tells you if the site has IPv6 forward entires, and doesn't let you know if you are connecting over IPv6 :( I wish it did, it would be cooler then.

[NewtonNet]

Ah, I didn't realise that. Are you sure though?

Does that mean that, say, if you have a client with v4-only access then it would still show an IPv6 address if the site you were visiting happened to have v6 AAAA records? Or is it the other way round - i.e. if you've got v6 access then it always shows a v6 address (assuming the site has AAAA records) even if you're connecting over v4?

Is there a test procedure to demonstrate that this is the case?

Mathew

[broquea]

I was thinking that you just disable IPv6, go to a dual-stacked website, see ShowIP display that it has ipv6 addresses associated with it, however that doesn't seem to be the case. I'd be pleasantly surprised if it now only displayed IPv6 for a website if it was connecting over it, but I don't remember that behavior before. I turned off IPv6 at home over my tunnel, and now only see the dual-stacked websites as IPv4.

[NewtonNet]

I'm pretty sure it does work 'as expected', although I must confess to not having really tested it as I always just assumed it did.

Perhaps when you last tested it there may have been a bug or some browser version incompatibility that was making it mispresent the address and implied transport? The plugin has been around for years (it started out as 'ipv6ident') and so I wouldn't be surprised if some quirks and misbehaviour haven't occured from time to time, particularly following the several major changes within Firefox.

Mathew

[jimb]

Wise words Jimb, and I think you're dead right - I shouldn't be noticing when I'm using IPv6 any more than anyone else notices they're using IPv4. Of course, like you, I also have that plugin (ShowIP?), so I do happen to notice and I must say I find it somewhat depressing (or worrying at least) how few IPv6-enabled sites there are out there, particular those that probably ought to be living and breathing the stuff (ISPs and the like).

Thanks for the pointers about what areas to look at studying next. I think that's partly my problem at the moment - IPv6 seems so vast[1] when you include the various transition methods, support functions, etc and to make matters worse many of them seem (to my untrained eye) to be competing with each other or, worse still, be obsolete ideas/concepts that are no longer in use!

Cheers,

Mathew

[1] I think this fact may well become one of the wakeup calls for the 'late adopters'... Indeed I've had colleagues at work ask 'which RFC [singular] covers IPv6?'...

Yeah not to mention that a lot of them are in flux, and many like DS-Lite are still just drafts, not even accepted RFCs yet.  But still good stuff and good to know.  DS-Lite is probably what Comcast will use to get customers on the internet when they run out of IPv4s, seeing that one of the main designers of it is from Comcast.  I actually like the approach.  What it does is set up a tunnel interface for IPv4 traffic, and tunnels it over IPv6 to a carrier grade NAT.  So your IPv4 traffic then goes to that CGN to get to the internet, and the IPv6 traffic goes native.  The coolest thing about DS-Lite is that since it's using a tunnel, the CGN can identify each user by the tunnel, allowing each user to have overlapping RFC1918 IPv4 addresses.  So for instance, 100 different customers can be using 192.168.1.0/24 and it doesn't matter, since the CGN can separate the traffic by the tunnel ID (e.g. the connection table will not only have original and translated source and dest IPs and ports, but also the origin tunnel ID).  Also, unlike TRT and NAT64/DNS64, no DNS ALG is needed to fake-up IPv6 addresses for IPv4 nodes.  So I think DS-Lite is going to be 'the winner' for ISPs.

[jimb]

I think broquea is right.  ShowIP basically shows the DNS, but AFAIK it has no idea what FF is actually using to connect.  I believe it just goes with the default behavior of FF and most OS, which is to prioritize in this order: global IPv6s, global IPv4, 6to4 & Teredo IPv6 addresses.

I know, for instance, that if I don't restart FF, that it still displays the IPv6 address of this forum even though I know it's connecting with V4.  But if I restart it, it figures it out.

[NewtonNet]

You know what? I think you're (both) right!

Worse still, it looks like this 'bug' was reported back in Aug '08!  :'(

http://code.google.com/p/firefox-showip/issues/detail?id=8

Oh well, I suppose it's still got its use(s).

Mathew

[jimb]

Yep.  For the most part it "works as expected", unless you pull the rug out from under it.