Re: draft-narten-ipv6-3177bis-48boundary-02.txt

[Mark Smith <ipng@69706e6720323030352d30312d31340a.nosense.org>]

On Fri, 14 Jul 2006 00:20:49 -0400 (EDT)
"Jason Schiller (schiller@uu.net)" <jason.schiller@mci.com> wrote:

> On Thu, 13 Jul 2006, Iljitsch van Beijnum wrote:
> 
> > It 
> > should be stressed that having different prefix sizes in the market  
> > place leads to higher operational cost because moving from one ISP to  
> > another then requires doing more work than just change a fixed number  
> > of high bits in all places where addresses appear in configurations.
> > 
> > Obviously it's easier to renumber from a small prefix into a larger  
> > one, but if different prefix sizes exist it can't be assumed that the  
> > only movement is from service providers giving out long ones to  
> > service providers giving out shorter ones.
> 
> If everyone accepts that ipv6 addresses have variable length subnet masks
> (and everyone has the appropriate tools to handle this) and the RIRs have
> a reasonably sound definition of what a good justification is for a given
> size IP block, then I fail to see the value of "classful" addresses.
> 

Go to fixed boundary allocations for end-sites, and _nobody_ needs
appropriate tools or processes to validate the sizes of requests from
end-sites. The bits we sacrifice in the IPv6 address space to do this
would be a cheap cost compared to the ongoing and continual costs of
managing variable length assignments to end-sites. If nothing else,
hopefully the RIR annual fees would go down, because they'd have less
work to do :-) Remember, there is at least 3/4 of the IPv6 address
space with nothing at all in it if we really do end up being completely
wrong. We've got plenty of IPv6 address bits to "waste" on addressing
convenience, as long as we don't blow up any route tables along the way.

> If a customer of ISP 1 adequately justified an ipv6 address with a
> particular size mask, and then the customer decides to instead purchase
> service from ISP 2, why wouldn't ISP 2 give out the same size ipv6 block?
> 

It seems to me that if you abstract that process a bit, the logical
conclusion is to just give nearly everybody the same prefix length, and
completely eliminate the administrative costs of the justification and
validation of the request process. Only the people dealing with the
public routing bits of IPv6 addresses have to understand and worry
about variable length prefix style address management.

I've understood the idea behind the /48 allocation was that it was
enough for nearly everybody - only the largest organisations or
entities would need shorter prefixes. This seemed to me to be
simplicity winning out.

I think introducing a second catagory of allocations, namely /56s, for
most smaller networks / homes etc, would address that "excessive"
criticism, yet still meet well enough a simplicity goal. It would still
increase the costs of address management though, as now allowance for,
and therefore pre-planning has to take place to cater for the instances
where the /56 needs to be increased to a /48. I think it is probably a
minimal enough cost to pay to address the "excessive" criticism,
although personally I'd prefer that even that address managment cost
wasn't incurred at all.

However, I think going to variable sized allocations for end-sites is
throwing the baby out with the bathwater. It increases addressing
administration costs, is much harder to understand for most network
administrators, and is much more error prone. I think IPv6 address
space is abundant enough and cheap enough that we can just eliminate
those IPv4-CIDR-style costs in the most common case, namely the very
large majority of end-sites, by just "throwing a bit of address space at
it". 

Regards,
Mark.