marcelo bagnulo braun allegedly wrote on 02 16 2009 2:47 AM:
> Hi Scott,
> 
> Scott Brim escribió:
>> A few thoughts on this quick draft:
>>
>> Pure higher-layer approaches aren't acceptable:
>>
>>   - We knew that already, for operational reasons other than
>>     renumbering, for example the shim6 argument. 
>>   
> 
> what is the shim6 argument?

The argument that took place between the IETF and network operators when
shim6 was recommended.

>>   - But that doesn't mean those technologies aren't essential, because
>>     even if an endpoint only has a single locator within a domain,
>>     both endpoints and networks will be multihomed and/or mobile, so
>>     they will at least appear to have multiple locators in global
>>     routing.  Shim6, HIP, SCTP, Multipath ... all are potentially very
>>     useful.  It's probably outside of the scope of RRG since they do
>>     nothing to solve routing scaling, 
> 
> what do you mean they do nothing to solve the routing scalability?
> If every site and host uses only PA addresses, the routing system would
> scale in the order of ISPs, reducing the contribution by the multihomed
> sites.

But that won't happen, because it requires site renumbering, multiple
locators per host within a domain, extra policy management and so on --
the things we discovered were operational showstoppers.  The host-based
mechanisms are vitally important but not as solutions to routing problems.


Scott

_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg

Reply via email to