[sfc] Re-reading draft-ietf-sfc-use-case-mobility

"Adrian Farrel" <adrian@olddog.co.uk> Thu, 19 January 2017 21:05 UTC

From: Adrian Farrel <adrian@olddog.co.uk>
To: draft-ietf-sfc-use-case-mobility@ietf.org
Date: Thu, 19 Jan 2017 21:05:01 -0000
Message-ID: <071801d27297$b3a48ee0$1aedaca0$@olddog.co.uk>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Thread-Index: AdJyl6vwRIf3N8rsQ6ea9YZzKTI/dA==
Content-Language: en-gb
Archived-At: <https://mailarchive.ietf.org/arch/msg/sfc/TULPLZwHQZbZ-Jwl_mBhh7qeMlU>
Cc: sfc@ietf.org
Subject: [sfc] Re-reading draft-ietf-sfc-use-case-mobility
Precedence: list
Reply-To: adrian@olddog.co.uk

Hi,

I've just been re-reading draft-ietf-sfc-use-case-mobility and am trying to draw
some conclusions about metadata.

I appreciate the work done to classify the metadata. This is helpful, but :-)

Looking at the cases for "synchronous injection" of metadata I'd like help
understanding whether this is:
- per SFC
- per SPI
- per flow
- per packet

I understand there is potential for overlap given the nested hierarchy implied.
I also understand that metadata may change periodically, and the change might
make it look like "per packet" when it is really "changing per flow".

But some clues would be really helpful.

Oh, one other thing. The end of 6.3 says "or carried with http header
enrichments within the user payload". If that option is chosen and melds with
the asynchronous option of "from the control plane environment by means of
individual standardized interfaces" does that mean that this use-case does not
have a requirement for NSH metadata. Just asking!

Thanks,
Adrian

[sfc] Re-reading draft-ietf-sfc-use-case-mobility Adrian Farrel