Hi Geoff:

Thanks for your help here. We certainly need some from all the WG at
this point.

This draft is a trade off between:
- compression coverage
- ease/size of code
- size of the dispatch/hc

As Julien commented in the list, we do not compress all types of
multicast addresses, and only compress matching scopes between source
and destination as follows:
- Global/ULA <--> Global/ULA
- Link local <--> Link local
- Link local  --> Multicast 
- unspecified --> Multicast

Also use 4 methods of compression:
- Stateful with 16 contexts
- Stateless based on MAC header
- Stateless based on well-known prefix (link local, default global)
- Stateless based on well-known patterns (multicast, zeroed out suffix
header)

Zach mentioned in the list that the stateful method might be too
expensive for some devices out there. In fact, the draft needs to
clarify which methods:
- MAY/SHOULD/MUST be supported for encoding
- MAY/SHOULD/MUST be supported for decoding.

The more cases we support, the more bits are required in the frame to
describe the encoding and the more code is necessary to encode/decode;
so we picked a trade-off based on existing experience. Clearly we need
more feedback from this group to feel whether the trade off is
acceptable or should be tuned, in particular from the angles of:

- ease of implementation. Are complexity and code size acceptable?
- coverage. Need more? Too much already?
- Make stateful compression 16 contexts a MAY? Make --> Multicast a MAY
or a SHOULD ?

Please provide your thoughts along these line on the list. Time is
running out.

Pascal

>-----Original Message-----
>From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On
Behalf Of Geoff Mulligan
>Sent: vendredi 10 octobre 2008 01:51
>To: 6lowpan
>Subject: [6lowpan] HC-01 draft
>
>Folks,
>  The draft of the new HC format has been published as a WG document:
>http://www.ietf.org/internet-drafts/draft-ietf-6lowpan-hc-01.txt
>
>There has been some good discussion around this, but would everyone
take
>up that task to read and review the current draft and provide feedback.
>
>Jonathan do you already have updates for the draft or is -01 the most
>current thoughts?
>
>So that we can close on this document sooner rather than later I am
>going to continue to bug the group to deal with this document.
>
>Please read it now and send comments to the list.
>
>       geoff
>
>
>_______________________________________________
>6lowpan mailing list
>[email protected]
>https://www.ietf.org/mailman/listinfo/6lowpan
_______________________________________________
6lowpan mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6lowpan

Reply via email to