Re: [Roll] Adoption draft-goyal-roll-p2p-measurement-01 as aROLL WG document

And then there is the option in most cost optimized environments that
one is happy with the P2P discovered routes as long as they are not
causing many retransmissions and the round-trip delay is acceptable.
The measurement mechanism is a useful optimization tool, but not
necessary for making P2P implementations work on a basic level.
Thus, they should be in separate docs.

Cheers,
  Anders
-----Original Message-----
From: roll-bounces@ietf.org [mailto:roll-bounces@ietf.org] On Behalf Of
Mukul Goyal
Sent: 13. april 2011 14:38
To: Philip Levis
Cc: ROLL WG
Subject: Re: [Roll] Adoption draft-goyal-roll-p2p-measurement-01 as
aROLL WG document

Phil

>I agree that currently the two seem tightly entwined, in that p2p needs
the measurement mechanism. But it might be that measurement is useful
independently of p2p, and so keeping their specifications separate will
prevent false coupling. It might be that once we are further along it
makes sense to merge them, but in my opinion keeping them separate for
now will lead to simpler and cleaner designs.

The measurement mechanism was originally in the P2P draft itself. It was
suggested to us that this mechanism should be its own draft precisely
for the reasons you have described above. However, it could be that Zach
is not referring to simply putting the two mechanisms together in the
same draft.

Thanks
Mukul

----- Original Message -----
From: "Philip Levis" <pal@cs.stanford.edu>
To: "Zach Shelby" <zach@sensinode.com>
Cc: "ROLL WG" <roll@ietf.org>
Sent: Wednesday, April 13, 2011 7:23:20 AM
Subject: Re: [Roll] Adoption draft-goyal-roll-p2p-measurement-01 as a
ROLL	WG document

On Apr 13, 2011, at 10:37 AM, Zach Shelby wrote:

> JP,
> 
> Although I agree that route metric measurement is necessary for p2p,
why in the world is this a separate document from the base p2p
specification? Furthermore there seems to be plenty of inefficiency in
creating a separate mechanism, and it should be explored how this could
be integrated into the same message exchange for setting up the p2p DAG.

> 
> Therefore I would be opposed to have a separate WG document for the
measurement mechanism, instead I think this work should be better
integrated into the base p2p specification. 

Zach,

I agree that currently the two seem tightly entwined, in that p2p needs
the measurement mechanism. But it might be that measurement is useful
independently of p2p, and so keeping their specifications separate will
prevent false coupling. It might be that once we are further along it
makes sense to merge them, but in my opinion keeping them separate for
now will lead to simpler and cleaner designs.

Phil

_______________________________________________
Roll mailing list
Roll@ietf.org
https://www.ietf.org/mailman/listinfo/roll
_______________________________________________
Roll mailing list
Roll@ietf.org
https://www.ietf.org/mailman/listinfo/roll