[BBLISA] anybody doing IPv6 for real operations?/possible presentation topic

[Internaut at Large]

Greetings,

On Tue, 2010-03-16 at 18:19 -0400, Dean Anderson wrote:
> On Tue, 16 Mar 2010, Internaut at Large wrote:
> 
> > Greetings,
> > 
> > On Mon, 2010-03-15 at 20:01 -0400, Dean Anderson wrote:
> > > I think the easiest way to play around is to install netbsd.  But I know
> > > of nothing quite turnkey at this minute.  Users come after ISPs on the
> > > install tree.
> > 
> > Right, us peons, the ones who want servers, who want IPv6, are the
> > ones who are the last to get what we want, right?  Thanks!  Dean this
> > isn't an ISP list, it's a USER list.  Think about your target audience
> > for the nonce?
> 
> Actually, its a large scale system administrators list.  I think
> inet-access is probably more oriented to users. (does that still exist?)

And ... large-scale systems administrators are usually ... that's right,
it starts with an "e" ... and ends with a "srs"  Do you need more hints?

> But it isn't about "getting what you want" nor about being //last// to
> get what you want.

Umm ... what else does "Users come after ISPs on the install tree" mean
then?  Apparently I'm too stupid to figure that out.

>   My point is about what can be had or given.  What
> you think can be had, just can't be had.

I've had it in the past, I don't think the problem space has changed
overmuch, it's still a problem, and there are still people working on
fixes, as opposed to trying to freeze things, and say "they (sort of)
work now, we aren't changing anything, and anyone who isn't in a good
position now, has to suffer for all eternity."

People like, I know your opinion of them, but ... NANOG, and other
folks.

>  Its not possible.  Its not
> possible because of route table size.

Actually, that was say a number of times, as the internet grew, for
various reasons. "There isn't enough bandwidth."  "Ethernet can only go
so far."  "Bang-addressing has gotten too complicated to keep track of."
and many other comments, that have seemed impossible at the time.  I
know you are simply an implementer, and see the limits of what you have
implemented, but those of us on the side of creating things tend to want
to solve problems, not just throw up our hands and give up.

>  Even with 2 million routes in the
> next generation of super routers, end users aren't going to get portable
> (ISP-independent) space, since every portable route must be in the
> global route table.

Why?  Seriously, why?  Because that's the way it's done now?  Do you
know why there are "next generation super-routers"?  Because we keep
getting to the next bottleneck, and solving that problem.  And getting
to the next one, and solving that problem.  So the 2-million routes
machines will work for a bit, as we grow into the next space, and
hopefully the 4 million route ones will not be too far behind, and them
maybe the 8 million route ones will manage to struggle through ...

> I mean think about it. Do you really suppose there will only be 2
> million end users, each with their own route? (much less)

Nope.  But not that many tomorrow.  The next day?  Well ... we'll see.

But it's going to be a pain for you, you are going to have to upgrade,
or not be able to serve your section of the world, right?  And you might
have to learn something ... new.

> As of November, there are currently 98836 blocks allocated worldwide by
> all 5 RIRs. (Regional Internet Registries). Right now, I've got 308298
> prefixes advertised for those 99,000 blocks.  So a route table of 2
> million can probably handle only 600,000 allocated blocks.

Right, which is more than 6 times the current size.  Don't you think it
might take a couple of years for us to double in size, much less eat 6
times our current size in blocks?  And by then, we might be able to
handle a route table of 4 million, thus doubling the space allocation to
1,200,000 blocks, unless a breakthrough happens, and they way they are
organized changes, so that, instead of simply up-scaling, we actually
change the way we are looking at things, in which case we might get an
order of magnitude change.  It's happened before. 

> The other question related to server operation, just a globally unique
> IP address (as opposed to an ISP-independent block), is a matter of cost
> and policy.

I think it's just policy at this point.

>  The server costs more to the ISP;

How does the server cost more?  Unless you are looking at all of the
bits themselves, moving the bits in the pipe shouldn't be any different
for you.  Unless you are hoping to limit us all to simple receive, like
a television, as opposed to exchange, like most of us "poor, last-served
end-users" want to do.  Bits are bits.  Upload, download, you shouldn't
care.

>  Hence the ISP charges
> more.

Sounds specious to me, but, granted the ISP charges more.  It also
charges more for a fixed IPv4 address, which is currently
understandable, because they are scarce, but, once they are plentiful
IPv6 addresses, a piece of your (I mean the ISPs) unearned revenue goes
away.  That seems to bother you.

> And those people who haven't paid (regular residential users) are
> going to find that they can't run servers by contract and technical
> implementation of that policy.

Because we are limited by paternalistic, and over-controlling ISPs?
Yes, and we will find a way to get around it, to run our servers, to go
back to being able to do all the things we used to be able to do, to
connect, to work, to communicate, to share.  To use the FTP and Telnet
equivalents of today that all this NATting took away from us.  Then the
ISPs contracted away from us, because they were the only game in town.

> Home residential users aren't going to
> be able to run servers, no matter what protocol they use.

Just, because you and the other ISPs say so?

> If you want IPV6 because you think it gives you servers at home, then
> you are just wrong, and you misunderstand the policy, administrative,
> and technical issues that are involved.

I understand the technical issues just fine.  You are right, though, I
don't understand the policy and administrative issues, even though I am
familiar with them.  As far as I can tell, they are simply arbitrary in
nature, and not actually mapping to anything in the real world, except
as a means of brow-beating someone who doesn't actually know the
technical issues, into line.  Perhaps, you can come down off your high
horse, and explain some of them to us?  Use little words please, because
apparently we aren't bright enough to understand them.

>  No insult intended, but that
> just isn't going to happen under any addressing scheme.

Because you won't let it?

Thanks.
> 		--Dean

-dkap, Wondering where "free as the internet" has gone to.

"When bad men combine, the good must associate; else they will fall one
by one, an unpitied sacrifice in a contemptible struggle." -Edmund Burke

"All that is necessary for evil to triumph is for good men to do
nothing." (Translated from Russian, from S. Bondarchuck's adaptation of)
-L. Tolstoy, "War and Peace"