I-D Action:draft-azinger-scalable-addressing-00.txt

[Brian E Carpenter]

Hi Gert,

On 2010-09-29 09:57, Gert Doering wrote:
> Hi,
> 
> On Wed, Sep 29, 2010 at 09:48:17AM +1300, Brian E Carpenter wrote:
>> To be clear, the classical model for scalable PA-based multihoming in IPv6
>> is and always was that a multihomed site would simultaneously operate a
>> different PA prefix from each of its ISPs and each host would simultaneously
>> have N addresses, one for each of those ISPs. Whether you like it or not,
>> this works perfectly and allows complete aggregation. 
> 
> I have serious doubt about the "works perfectly" part.
> 
> The idea is very nice, in theory (and I do like the shim6 approach).
> 
> Have the practical problems "how do I find the combination of source + 
> destination IP out of a set of N possible sources and M destination 
> addresses that works 'best' for me" (with locally varying metrics for
> what is 'better' - latency, throughput, price, firewall policies, 
> upstream RPF) been solved by now?
> 
> (This is partly rhetorical question, and partly real curiosity: that always
> seemed to me the hard part, and I really wonder whether I missed some 
> new developments here that would make this applicable for companies like
> Petra's, with 5 uplinks, strong requirements for certain destinations to
> be reached via a specific uplink, firewalls all over the place, etc.)

Shim6 does it by probing with REAP if the initial path fails. Choosing
the "best" path initially is indeed a hard problem.

    Brian