Question about pshim6

[Sébastien Barré <sebastien.barre@uclouvain.be>]

Hi,

It may have been discussed during the meeting but I could not attend 
unfortunately.
I have a question regarding this paragraph from section 3 of the draft :
---
When H1 sends the first packet addressed to the CMULA of H2, the
 packet is intercepted and processed by the P- Shim6 box of the
 multihomed site.  The P-Shim6 box retains the data packet and
 initiates the 4-way exchange to create a shim6 context with the
 P-shim6 box of the peer network.  This exchange uses the PA addresses
 as locators and the CMULAs as ULIDs.  Once that the shim6 context is
 established between the local P-shim6 box and the remote P- Shim6
 box, the local P-Shim6 box can forward the data packet with a shim6
 payload header, referring to the established shim6 context.
---

I wonder why it would not be possible and more efficient to send the 
first data packet as soon as we know one locator for that packet.
That is, I think we could use the same approach as host-based Shim6, in 
that the pshim6 box could perform the following steps :
- the proxy of host H1 receives the first packet from H1 to H2. With the 
CMULA of H2 as destination address.
- The proxy gets the corresponding locators, either from its cache or a 
reverse DNS lookup.
- The proxy chooses one of the obtained locators as destination for the 
data packet, and immediately sends it. We thus send the packet 2RTTs 
earlier than in the current pshim6 approach.
- The proxy applies some heuristic to decide if it initiates a Shim6 
negotiation, in the same fashion as host-based shim6.

Actually that approach is similar to what is proposed in the second 
option of section 7 (support for legacy sites and hosts). Indeed, what I 
propose is to behave as if the corresponding node were legacy, until we 
find it useful to use Shim6. Also, even if we decide to immediately 
start the shim6 negotiation, we can still do it in parallel with the 
data transfer, rather then before the data transfer.

Also note (if my understanding is correct) that in the case that both 
hosts in a communication use CMULAs, than the Shim6 negotiation is no 
longer required for providing transport layer survivability (There is 
only one CMULA candidate for rewriting an arriving packet, thus we do 
not need to ensure that one precise locator is used as ULID). OTOH, 
Shim6 is still useful for benefiting from the REAP protocol.

Am I missing something ?

regards,

Sébastien Barré.

--
Sébastien Barré
Researcher,
CSE department, UCLouvain, Belgium
http://inl.info.ucl.ac.be/sbarre