[ppml] IPv6 flawed?

Iljitsch van Beijnum iljitsch at muada.com
Mon Sep 3 04:57:21 EDT 2007

On 3-sep-2007, at 9:14, Owen DeLong wrote:

>> IETF is still working... And it's not like adopting PI in the mean
>> time helps.

> I strongly disagree.  Adopting PI in the mean time can serve as very
> useful input to express to IETF that they _MUST_ solve the routing
> problem in a scalable way.  And ID-Locator split is long overdue.

That's like everyone buying SUVs to signal that this global warming  
issue _MUST_ be solved.

My interpretation of the adoption of IPv6 PI was "the operational  
community doesn't consider routing table size an important issue".

> Shim6 is far too complicated to actually see reliable implementation
> in the average enterprise

Did you ever read the IPsec or SNMP RFCs? Those are at least an order  
of magnitude more complex, and enterprises managed to implement/ 
deploy them just fine.

> and it does nothing to help with easy renumbering.

My wireless keyboard doesn't help with easy renumbering either. I  
guess this means it's not a good keyboard.

> PI is a necessity.  CIDR and PA were horrible hacks

Without aggregation, routing only works by coincidence. Even WITH PA  
30k ASes results in a 230k routing table...

> put in
> place as a band-aid until IPv6 could be engineered to truly
> solve the routing problem in a scalable fashion.

Engineering scalable routing is not hard: create a big, fat  
hierarchy. It's getting scalable routing deployed and breaking pretty  
much all ISP business models in the process that is a challenge.

> I have no
> idea why an ID/Locator split was not made a built-in part
> of IPv6, but, I believe it can be added without reworking the
> protocol, so, there is still a way forward.

An id/loc split is certainly interesting, but what it basically does  
is say "having stable FQDNs and renumbering IP addresses is too hard,  
so now we're going to have stable FQDNs that map to stable id  
addresses which map to loc addresses which we WILL be able to  
renumber". It would be helpful to see evidence of that last part  
before we go down this road.

> IPv6 has worse downsides than IPv4.  I can see a few rare instances
> where autoconfiguraiton could be considered a useful feature, but,
> even in those cases, I don't see where it is significantly better than
> DHCPv4.

At the last IETF meeting, the DHCP server flaked out several times.  
IPv6 stateless autoconfig was solid as a rock. Same thing at a RIPE  
meeting a few years ago, unless I misremember. DHCP is actually not  
that easy to get right in the presence of failing hard- and software.  
With stateless autoconfig you can boot up a new router and take the  
old one down a few minutes later with no real service interruption.

More information about the ARIN-PPML mailing list