Re: [Mip6] draft-dupont-mip6-dhaadharmful-00.txt

[Francis Dupont <[e-mail]>]

 In your previous mail you wrote:

   > => there is no such "load condition" defined in RFC 3775...
   
   RFC3775 says:
   
   	"...the sending home agent MAY
         dynamically set the Home Agent Preference value, for example
         basing it on the number of mobile nodes it is currently serving or
         on its remaining resources for serving additional mobile nodes;
   	such dynamic settings are beyond the scope of this document."
   
=> there is a large difference between a parameter which MAY be derived
from load condition and a load condition field in a message.
   
   > => DNS has caching so is more efficient in term of delays.
   > For the signaling DNS and ICMP are at the same order.
   
   I think this is a misunderstanding. I didn't mean the signaling delay
   for DNS/DHAAD request/reply. I agree that they should be in the same
   order.
   I meant that delay and signaling overhead can be reduced if
   the HA discovery/assignment is intergrated in IKE exchange that is
   needed for bootstrapping anyway, e.g., the anycast IKE_SA_INIT is
   replied by 
   the HA with the lowest load (would require co-ordination between HAs)
   
=> so you compare with a potential mechanism about we have no details...
I still believe caching will always give an advantage to DNS, i.e.,
the DNS cache should be closer than the HA in the common case.

   I think Alex had the same comment. I think we had a different
   understanding of the term HA-switch.

=> for me HA-switch is the mechanism described in
draft-haley-mip6-ha-switch-00.txt. Obviously a HA can't crash
and after send HA-switch messages.

   trigger the HA-switch mechanism (i.e., bootstrap with new HA).

=> this is HA reselection.
   
Regards

[e-mail]

_______________________________________________
Mip6 mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/mip6