Re: [RRG] IPv4 shortage, new features and IPv6 inevitability

[Iljitsch van Beijnum <iljitsch@muada.com>]

On 22 nov 2007, at 4:59, Robin Whittle wrote:

> I think the only solution to address depletion is to divide the
> remaining free space and currently allocated but underutilized space
> into smaller blocks than has been the case (longer prefixes), down
> to /30 or whatever.

This completely ignores the current reality that the average address  
block given out is almost 30,000 addresses. 90% of the address space  
given out is very large blocks going to ISPs, who presumably need it  
for their access services. If you need a million addresses you could  
also take 16 /16 blocks or maybe even 256 /24 blocks, but certainly  
not a quarter million /30s. Also, IF IPv4 address space is available  
it's going to be in blocks significantly larger than that, not only  
because the smallest that can be used today is /24, but also because  
the vast majority of the presumably reclaimable address space is in  
the form of /8s (and almost all of the rest in /16s). Some slicing and  
dicing may be required, but from a /8 down to a /30? Don't think so.

> With the existing reliance on BGP alone, that
> would lead to a completely unsustainable explosion in the number of
> prefixes.

An interesting thing about BGP is that if you break it, you see that  
immediately. That means you'll revert back to the previous  
configuration within minutes. So anything that has enough impact to  
really break BGP simply won't be deployed.

> I think the continuing need for IPv4 address space, including
> smaller chunks of it with greater use of NAT, will continue for the
> foreseeable future - 5 to 10 years at least.  Having an IPv6 address
> does not alter the need of the vast majority of users for direct (or
> via NAT) access to IPv4.

It's interesting to contrast the situation the big content players are  
in versus the one ISPs are in. As a content network, you need very  
little IPv4 space. Moving to IPv6 is pretty much a binary thing: you  
either enable it or you don't. So for content people enabling IPv6 is  
hard and sticking to IPv4 is easy. ISPs on the other hand needs lots  
of fresh IPv4 space to connect new customers, but they are in the  
position to roll out new deployments for new users and keep existing  
customers on existing stuff so they don't have to go to IPv6 wholesale  
at once. So for ISPs sticking to IPv4 is going to be problematic due  
to the depletion while introducing IPv6 is relatively easy.

So we'll probably see a world where more and more eyeballs are on IPv6  
(which means in practice either IPv6-only with proxies/NAT-PT or  
IPv6+IPv4/NAT dual stack) while most of the content will stick to IPv4- 
only for a much longer time.

Any claims of 5 or 10 years or even longer is highly suspect because I  
don't see a quantitative basis for it. My best guesstimate about the  
IPv4 depletion is that we'll run out in 2012, so yes, in 5 years most  
people will still be running IPv4. But if good transition mechanisms  
are available, it will be easy to move to IPv6-only while still having  
access to IPv4 resources and then a significant number of people could  
be living on an IPv6-only network fairly soon.

> Iljitsch, despite what you wrote and what quite a few people believe
> about IPv6 and about IPv4 utilization, still I believe (for all the
> reasons I stated) that IPv6 offers no short to medium term (1 to 5
> years) benefit to ordinary Internet users (and therefore their ISPs)
> compared to the costs of adopting it.

The benefit is that you get to connect new users to the network for  
more than just the next 5 years.

Building a house also doesn't give you an immediate benefit: it costs  
a lot of money while building but you can't start living in it until  
you're done. But houses are built every day.

The costs of adopting IPv6 are quite small, especially if you take  
enough time to do it.

> I still believe that for the next 5 to 15 years most users will find
> it better to squeeze more usage out of IPv4 address space

That only prolongs the pain. Worst case would be that address policies  
become very strict and people spend enormous amounts of time and money  
getting a few IPv4 address but we don't really "run out" so IPv6  
deployment doesn't happen. We need fresh IP addresses to be available  
one way or the other, without that, we'll all be wasting a lot of time  
and money on addressing that could have gone to something productive.

> > > 1 - IPv4 address depletion is the most urgent architectural
> > > problem facing the Internet - and far better recognised than
> > > BGP stability and router scaling problems with the growth of
> > > advertised prefixes.

> > Ah, but that's a solved problem. RRG = research, IETF =
> > engineering. IPv4 depletion = operation. We all know what needs
> > to be done here. A wise man once said: just do it.

> I see no consensus at all on what needs to be done regarding IPv4
> address depletion.

> I think you imply that we all agree the answer is to move to IPv6.

Knowing that something needs to be done and agreeing to do it aren't  
the same thing.

> The trouble is, as I outlined above, IPv6 only solves the IPv4
> address depletion problem once everyone - or almost everyone - has
> moved to IPv6.

That's why we still need to work on transition mechanisms in the IETF.  
Dual stack isn't it.

> I think it was a mistake for IPv6 to have such long
> addresses, adding to the length of every packet.  64 bits should
> have been fine

That wouldn't have made any difference to our deployment issues, it  
would just have meant no CGA/HBA or stateless autoconfig. The real  
answer is of course that addresses need to be variable length.

> > I'm not convinced. The issue isn't the number of places that need
> > an address block, but the number of places that need an
> > individual address.

> I think it is both, although NAT tends to reduce the quantity of
> addresses each end-user network needs.

You assume a model where it would be possible for every device that  
connects to the network to get an individual public address. That's  
simply not possible with IPv4, because the need for addresses is so  
dynamic that ISPs can't realistically provision addresses this way.  
(For instance, currently I don't need any addresses at home. When I  
get home and start copying some data between different computers I  
need 4 or 5 addresses.)

The current model is that in most cases, an ISP provisions one address  
per line to a customer. If this changes to a larger number of lines  
per address this makes NAT more problematic for end-users, but yes,  
that could save a lot of address space.

> > > 5 - IPv4 address space utilisation could easily be improved if
> > > there were suitable policies and slicing and dicing
> > > technologies. Ping responsive host rates in advertised space
> > > are around 4%:

> > Meaningless. First of all, they also pinged unrouted space.

> It is not meaningless.  My survey and the much better work at USC ISI:

>  http://www.isi.edu/ant/address/
>  http://www.firstpr.com.au/ip/host-density-per-prefix/

> both surveyed address space which was advertised in BGP.

I asked the guy from ANT, he said that they also pinged unrouted space.

> > > So there is plenty of room for improvement.

> > No. Any effort spent on getting back IPv4 space for new uses is
> > wasted effort, because we need to move to IPv6 in the slightly
> > longer run anyway.

> Geoff Huston estimates 5 to 20% utilization.

Could be. I'm pretty sure that many ISPs never bother to reuse old  
address space when customers leave etc but simply request new stuff.

The real question is: why is the utilization so low? And is that  
reason still present today? If so, what would we need to do to  
overcome that factor? And if we do that, how much extra IPv4 time  
would that give us?

We already see that ARIN doesn't want to reclaim any of the legacy  
class A address space, even though that only means 40 instances of  
administrative work for 650 million IPv4 addresses - a pretty sweet  
deal compared to pretty much any other effort that you can think of.

> IPv6 is years - probably decades - away from being sufficiently
> widely and robustly deployed in applications and OSes.

Nonsense. I'll prove it to you by turning off IPv4 right now on my  
system and sending you this message running IPv6-only using only  
software that came with my Mac and a dual stack server elsewhere on  
the network.

I'm not saying it's possible to turn off IPv4 today without any impact  
on functionality (I can't print or use my instant messaging client  
right now for instance) but we've come a long way since when I started  
with IPv6 and I could only do ping6 and traceroute6.

> But why would I, or most other end-users (and their ISPs) get IPv6
> connectivity, ensure our most important devices, hosts OSes and
> applications are IPv6 compatible, when there is no benefit now or in
> the next few years over continuing to use IPv4 and leaving my
> computer undisturbed?

Forget end-users. They don't know about this stuff. ISPs will start at  
some point but they're going to wait some more because if you have the  
choice between spending money now or spending it tomorrow, the latter  
is usually a no-brainer.


--
to unsubscribe send a message to rrg-request@psg.com with the
word 'unsubscribe' in a single line as the message text body.
archive: <http://psg.com/lists/rrg/> & ftp://psg.com/pub/lists/rrg