[Dime] discuss encoding of bulk commands /RE: very drafty IETF#83 meeting minutes uploaded

Thanks for the meeting minutes. 

As the notes refer to a discussion about different means to signal bulk operation,
let's go through this exercise and evaluate the three candidates to signal
bulk commands.

Just to summarize the three alternatives according to Mark's presentation:
(a) New group commands
(b) Re-use existing commands and command codes, indicate group operation
on the attached context (AVPs) within the message, e.g. by adding group
identifier, flags, ..
(c) Single new bulk command (1 new command code), indicate Diameter command,
to which the group operation applies, within the message

Whereas I see (a) as straight forward method, a counterargument could be the
need of additional command codes for each command, which should be enabled
for group operation. Other SDOs, which specify new command codes, need
to follow the same rules and request new command codes for all existing commands
which should be bulk-operation enabled.

With (b), all existing messages could be re-used and potentially bulk operation enabled
by using a different application ID and adding an additional label to the message, e.g.
flags, Group session AVP, etc. Such approach allows much flexibil8ity, but we need
to ensure  compatibility with proxies etc, which are not bulk-operation enabled.
Possibly using a new application ID solves that issue already.

With (c), the DIME WG could specify a single dedicated command for bulk operations,
and the actual payload carries the command code of the Diameter message to be
executed as bulk operation (e.g. RAR, STR, ..) as well as the associated command's
AVP, plus a group identifier. Any existing and new command, which is so far used
for single session operations, can be bulk operation enabled. Formatting of such
bulk operation message is still open, maybe it has impact to existing applications'
command parser. If we find a suitable encoding scheme for such container message,
it's a concept that can be easily adopted by any Diameter command and application.

We'd be interested in more opinions about the different approaches and I hope that
this eMail can trigger such discussion.

marco

> -----Original Message-----
> From: dime-bounces@ietf.org [mailto:dime-bounces@ietf.org] On Behalf Of
> jouni korhonen
> Sent: Dienstag, 10. April 2012 11:53
> To: dime@ietf.org
> Subject: [Dime] very drafty IETF#83 meeting minutes uploaded
> 
> Folks,
> 
> They are up now for public review. Thanks to Shwetha for taking excellent
> minutes.
> 
> - Jouni
> _______________________________________________
> DiME mailing list
> DiME@ietf.org
> https://www.ietf.org/mailman/listinfo/dime