Hello Anand:

2 good points , that we need to continue in different threads if we do.

The source routing optimization done by dropping the addresses on the way
certainly has benefits. However, there could be loss of a natural "symmetric"
property that one might want to enforce between communicating end nodes in the
routing path. By symmetry I mean using the same path in both the directions of
communication. Policy based routing, centralized routing, for instance, could
be potential users of this property. May be this does not represent a common
use case. But nevertheless, we have to be aware of this side effect which 
RFC6554
swapping process does not have.


?  Pascal: Yes, Simon made that point yesterday.

?  It is generally not a good idea to reverse a routing header. The RH may have 
been used to stay away from the shortest path for some reason that is only 
valid on the way in (segment routing).

?  P2P RPL reverses a path that is learnt reactively, so we have a real 
protocol for doing that sort of thing as opposed to an echo.

?  Reversing a header is discouraged by RFC 2460 (for RH0) unless it is 
authenticated (which means AH). We do not authenticate the RH3, there are a 
number of reasons for that, general deprecation of AH, no use of AH in LLNs 
etc... Note that AH does not protect the RH on the way, it is just a validation 
at the receiver for the sole value of reversing it.

?  A RPL domain is usually protected by L2 security and that secures both RPL 
itself and the RH in packets, at every hop

?  The benefit of saving energy and lowering the chances of loss are seen as 
overwhelming compared to the value of possibly reversing the header

?  Yet we might define a variation where we do not pop out the first entry as 
we go. I do not see a consensus on the value of doing it now.



Going further, the DAO projection proposal
(draft-thubert-roll-dao-projection-02.txt) will have several virtual roots
inside the RPL domain. The automatic assumption of a well known root may not 
apply
when nodes within RPL domain communicate with each other. I suppose it will
have a bearing on the RH3-6LoRH performance.


?  It is not really an assumption, but something we leverage as we go. The RPI 
is very much like a context indicator. If an address shares a lon prefix with a 
root, adding the RPI of that root is actually a compression technique.


The above observations are not serious, but feels good to ponder over. Will be
happy to receive your comments.

Thanks Anand!

Pascal





On Monday 18 January 2016 11:24 PM, Pascal Thubert (pthubert) wrote:
>
> Dear all
>
>
>
> The picture below illustrates how the RH3 6LoRH works with draft 03 in a case 
> like Root -> A -> B -> C -> leaf
>
>
>
>
>
> The first 6LoRH is expected to be a full address (128 bits) to set up a 
> reference and the next 6LoRH are expected to be smaller and just override the 
> rightmost bits which form the delta from the reference.
>
>
>
> Proposal: we could consider that the 128bits source of the IP header before 
> the RH3 is the reference to start with.
>
>
>
> With that even the first hop could be compressed the same way as the other 
> hops. With RPL, the root is the encapsulator if IP in IP in used. Good thing, 
> in that case the root is totally elided with the IP-in-IP 6LoRH.
>
>
>
> So this simple proposal saves up to 16 octets (thats in the case with a 
> single subnet and all addresses differ only by the last 2 bytes). Im willing 
> to add it in the next revision.
>
>
>
> Any opposition?
>
>
>
> Pascal
>
>
>
> --
> This message has been scanned for viruses and
> dangerous content by MailScanner, and is
> believed to be clean.
>
> _______________________________________________
> 6tisch mailing list
> [email protected]<mailto:[email protected]>
> https://www.ietf.org/mailman/listinfo/6tisch



--
This message has been scanned for viruses and
dangerous content by MailScanner<http://www.mailscanner.info/>, and is
believed to be clean.
_______________________________________________
6tisch mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6tisch

Reply via email to