Re: Discussion on Charter

More on Fred's astute comments.

A lot of the QoS issues -- which seem to take up a good share of the ink
-- are layer 2 issues.  That puts them outside the scope of IETF as a
whole as layer 1/2 standards are not addressed by IETF (a decision, btw,
that I agree with).  

For example, if you want GETS type service, one translation is a need
for assured access at layer 2.  With contention queues (like 802.11
WiFi), you simply can't get that.  And nothing you do at layer 3 and up
will change that fact.  If, on the other hand, you have a scheduling
queue (802.16, .20, .22), you can provide layer 2 QoS.  At this point,
we do have an issue -- how do you control that stuff from higher layers.
Here the otherwise-healthy division between IETF and IEEE needs a
connection.  This need has been often recognized but never really hit
stride ... perhaps when WiMax stuff arrives in the marketplace in force
that will change (why build a control mechanism when there's nothing to
control?).






On Wed, 2005-02-23 at 15:11 -0800, Fred Baker wrote:
> At 09:48 AM 02/23/05 +0000, Ian Brown wrote:
> >Pratik - this has been a long-running debate. But it does seem logical to 
> >have the experts on a particular protocol (the responsible Working Group) 
> >be involved in any updates that are required to provide these 
> >capabilities. Given that the WGs conduct the vast majority of their work 
> >by e-mail, and meet at the physical IETF meetings three times a year, 
> >would it be particularly more efficient to try to gather all of this 
> >expertise in one place, whether a rechartered IEPREP or a new WG?
> 
>  From my perspective, it might be.
> 
> Related work is going on in a variety of places right now. Some relates to 
> RSVP, to MPLS, to diffserv code points, to SIP, and IIRC a few other 
> things. The RSVP and diffserv WGs are right now dormant, and the other 
> working groups find this rather peripheral to their central foci. If they 
> can each severally sit down and get something done, fine. But consider:
> 
>   - the SIP work was originally proposed in Adeliade, five years ago
>   - the RSVP work builds on work that was finished in 1998, and the WG has
>     been closed since then
>   - NSIS work, if one chooses that route, is proposed, but at this point very
>     sketchy
>   - Diffserv has likewise been closed for at least two years
>   - We have discussed parts of it with TSVWG, but it is just a few weeks ago
>     that they decided to even consider responding to it by allowing the
>     documents to become WG drafts
> 
> So putting together a WG - whether ieprep with an appropriately changed 
> charter or something new - that can focus on this problem and put out a 
> solution seems more likely to actually produce a result within our lifetime 
> than waiting for WGs that don't exist or don't care to decide it is important.
> 
> YMMV 
> 
> _______________________________________________
> Ieprep mailing list
> [email protected]
> https://www1.ietf.org/mailman/listinfo/ieprep
-- 
b