A question about the role of the IESG

Jonne.Soininen at nokia.com Jonne.Soininen at nokia.com
Wed Jan 8 08:39:12 CET 2003


Hi Brian,

this was exactly what I was looking for. Being in the IETF only for four/five years,  hasn't provided me, yet, with enough information about the past to understand the history behind certain decisions. 

Anyways, I wonder if some of the decisions or assumptions that were done in 1992, which is over ten years from now need a bit of reconsidering. I guess time will show when we get a bit further with this work.

Cheers,

Jonne.

> -----Original Message-----
> From: ext Brian E Carpenter [mailto:brian at hursley.ibm.com]
> Sent: Wednesday, January 08, 2003 7:57 AM
> To: Soininen Jonne (NET/MtView)
> Cc: problem-statement at alvestrand.no
> Subject: Re: A question about the role of the IESG
> 
> 
> Jonne,
> 
> Hmm. I thought it was clear. Let my try some more words...
> 
> Jonne.Soininen at nokia.com wrote:
> ...
> > 
> > > As far as the IETF goes, the authority used to belong to 
> the IAB, and
> > > was given to the IAB by the people who funded the ARPANET and
> > > the early
> > > Internet. Then there was a popular rebellion in 1992, and the IETF
> > > decided to give its authority to the IESG as a result. 
> That's when the
> > > current nominating process for the IESG and IAB was 
> invented; it's the
> > > IETF community that puts the IESG in place, and agrees to 
> accept its
> > > authority.
> > >
> > > I don't see how you can have an SDO without a management 
> structure.
> > > Ours happens to be our very own IESG, but we also run with a very
> > > light staff, which is not the case in the ETSI world.
> > 
> > I guess, I really struggle here trying to make my point. I 
> not trying to compare IETF to any other SDO as such, and not 
> trying to criticize the need for a management/coordinating 
> structure. What I have noticed is that IESG has _relatively_ 
> much power over the IETF standardization process, 
> 
> Yes. The IETF gave it that power by its usual consensus process, back
> in 1992. 
> 
> > _relatively_ to the power IESG reports little to the community, 
> 
> We could argue that for ever, but let's take it as input to the
> problem statement that this group is supposed to write.
> 
> > and the IESG is not directly *elected*, but rather 
> *selected* by rather unique process. 
> 
> Because the IETF has no defined membership, elections are impossible
> (just as voting in WGs is impossible). So the current process was
> invented as the best approach people could invent to a 
> consensus-driven
> selection process. It isn't intended to put the most popular people
> in the IESG and the IAB. It is intended to put the best 
> possible people
> there, according to consensus among a representative 
> nominating committee.
> 
> > What I am trying to understand - personally - here is how 
> was the NomCom process created (opposed to e.g. elections, 
> ISOC nominations, or anything else), 
> 
> By the "poised" WG, using normal IETF WG process. You can get 
> some hints
> at http://www.ietf.org/html.charters/OLD/poised-charter.html
> 
> > why has the IESG been given the role of the guardian of the 
> IETF process and the IETF output (vs. e.g. a mere 
> coordinating role between areas/WGs), 
> 
> Because with no voting being possible in WGs or the IETF as a whole,
> the power of decision *must* be given to somebody, and the choice was
> to give it to the IESG (previously the IAB).
> 
> > and why wasn't reporting seen that necessary.
> 
> I don't think it was an issue in 1992. The IETF was smaller.
> 
> > 
> > (I know there are certain efforts to enhance the reporting 
> part, e.g. ID Tracker. However, this is rather recent, and 
> IESG minutes are still just a list of results rather than 
> something comparable to for instance the WG minutes.)
> 
> Correct. Those minutes are a record of decisions taken. The 
> IESG has never
> minuted its internal debates, partly to retain collegiality and partly
> to avoid possible legal liability, as I understand it.
> 
> Hope this is clearer
> 
>    Brian
> 
> > 
> > Cheers,
> > 
> > Jonne.
> > 
> > >
> > >    Brian
> > >
> > > Jonne.Soininen at nokia.com wrote:
> > > >
> > > > Hi Dave,
> > > >
> > > > sorry about the long delay in answering this mail. However,
> > > I guess I did not explain myself correctly in my first mail.
> > > I do have read at least some assorted documents from the
> > > ietf.org, but seems from your mail that I would have not read
> > > the correct ones... ;)
> > > >
> > > > My experience (obviously not the same as yours) are from
> > > the IETF, and 3GPP. There are many differences in these two
> > > organizations, and in their purpose. So, most probably one to
> > > one comparison would not be of much use as what works for one
> > > may not work for the other. So, I will not even try. However,
> > > I'll explain here where my comment came from:
> > > >
> > > > In 3GPP, there is no appointed management like IESG, and
> > > IAB in the IETF. There is hierarchy, though. The work is done
> > > in two levels in Technical Specification Groups (TSGs) that
> > > coordinate Working Groups (WGs) Thus, the TSGs in some sense
> > > could be compared in the hierarchy to the level of IESG in
> > > the IETF. The main difference (in my mind) is that the TSGs
> > > are practically working groups. They are open to the whole
> > > 3GPP community. Anybody from 3GPP community interested in the
> > > work of a TSG can go to their meetings a affect the
> > > decisions. Both the TSGs and the WGs have a chairperson and
> > > one or more vice chairs. These are elected by the TSG or the
> > > WG in question - they are not appointed.
> > > >
> > > > This was what I used in my mind as a comparison point. I do
> > > not want to start a huge comparison between the IETF, and
> > > 3GPP here. I do not think that brings us much. We would just
> > > end up in a rat hole about the philosophical differences of
> > > the organizations, which would not bring us any closer to the
> > > problems that we face in the IETF now. What I wanted to
> > > understand was, how did IETF end up with this structure that
> > > we have now. Though that IETF is an open organization some of
> > > the current procedures seem to be in conflict with the
> > > openness principle - at least to the untrained eye. Most
> > > probably there was a reason for this at the time of creation.
> > > I would just like to know what that reason was.
> > > >
> > > > Cheers,
> > > >
> > > > Jonne.
> > > >
> > > > PS. Thanks for your article. However, I do not think it
> > > deals with the same issue that I was trying to raise.
> > > >
> > > > > -----Original Message-----
> > > > > From: ext Dave Crocker [mailto:dhc at dcrocker.net]
> > > > > Sent: Friday, December 20, 2002 8:23 PM
> > > > > To: Soininen Jonne (NET/MtView)
> > > > > Cc: problem-statement at alvestrand.no
> > > > > Subject: Re: A question about the role of the IESG
> > > > >
> > > > >
> > > > > Jonne,
> > > > >
> > > > > You might want to carefully read assorted documents 
> available via
> > > > > <http://www.ietf.org, before making the sort of sweeping
> > > > > assessment you
> > > > > offer.
> > > > >
> > > > > As to your question about choice, one place you might look is:
> > > > >
> > > > >          
> <http://www.brandenburg.com/ietf/ietf-stds.html#StdWay30>
> > > > >
> > > > > The rest of that article may also be helpful.
> > > > >
> > > > > d/
> > > > >
> > > > > Friday, December 20, 2002, 6:05:12 PM, you wrote:
> > > > > Jonne> According the current processes, and practices in the
> > > > > IETF it seems that the IESG has almost unlimited power over
> > > > > the IETF. The IESG is not really accountable to the IETF
> > > > > community (e.g.
> > > > > Jonne> meeting minutes are not public and only the collective
> > > > > decisions are communicated), the IESG has no responsibility
> > > > > towards the general public of the IETF, and it is seen as the
> > > > > management of
> > > > > Jonne> the IETF. In the current model, the IESG supervises,
> > > > > and has control over the work of the community, and the WGs
> > > > > execute the operative work. This is somewhat parallel to the
> > > > > model of a company
> > > > > Jonne> where the responsibility is always upwards. In
> > > > > general, this is rather unique arrangement in a community
> > > environment.
> > > > >
> > > > > Jonne> I was just wondering, how this operative model was
> > > > > chosen, and why this is seen as a more effective model than
> > > > > models used in other SDOs where the community steers itself
> > > > > without hierarchical
> > > > > Jonne> management?
> > > > >
> > > > >
> > > > > d/
> > > > > --
> > > > >  Dave <mailto:dhc at dcrocker.net>
> > > > >  Brandenburg InternetWorking <http://www.brandenburg.com>
> > > > >  t +1.408.246.8253; f +1.408.850.1850
> > >
> 
> -- 
> - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
> Brian E Carpenter 
> Distinguished Engineer, Internet Standards & Technology, IBM 
> On assignment at the IBM Zurich Laboratory, Switzerland
> 


More information about the Problem-statement mailing list