[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Updated reorg proposal, based on Monday's discussion
- To: iesg@ietf.org
- Subject: Updated reorg proposal, based on Monday's discussion
- From: Harald Tveit Alvestrand <harald@alvestrand.no>
- Date: Wed, 06 Aug 2003 18:18:21 -0700
Some changes, some clarifications, some points marked <UNCERTAIN>.
Added a section of "problems" at the end, which is not for inclusion in the
final version, but for not losing track of what people have mentioned while
discussing it.
What do people think?
Harald
(BTW, the "is intended to address" section was EASY to write.....)
------------------------------------------------------------------
A reorg plan for IETF decision-making
--------------------------------------
This plan is intended to address the following issues in
draft-ietf-problem-issues-02:
2.6 The IETF Management Structure is not Matched to the Current Size and
Complexity of the IETF
in particular the subsections
2.6.1 Span of Authority
2.6.2 Workload of the IESG
2.6.3 Procedural Blockages
2.6.4 Consequences of Low Throughput in IESG
2.6.6 Concentration of Influence in Too Few Hands
It will also have an effect on the following issues:
2.1: Participants in the IETF do not have a Common Understanding of its
Mission (helps, by causing more of the strategy for the areas to be
documented)
2.3 The IETF has Difficulty Handling Large and/or Complex Problems
(helps, by causing more people to be responsible for the "big picture
view")
2.6.5 Avoidance of Procedural Ossification (harms; since it increases the
need for written guidelines, it increases the need for vigilance in
guarding against ossification)
2.6.7 Excessive Reliance on Personal Relationships (helps, by putting more
people in close contact with others through formal roles)
Outline
-------
The following changes are made to the function of the IESG:
- The number of people in the leadership is EXPANDED from the current 13
to a number somewhere between 25 and 40.
- Each member of the decision-making group is a member of TWO kinds of
subgroups:
- An "area council", charged with overseeing the management of the
document development process and the working groups
- A "review team", charged with approval of documents for one area.
- Each "area council" is headed by one "area supervisor". The set of "area
supervisors" together form a "technical leadership team".
- Each "review team" is headed by the "area supervisor" for the area, and
consists of one member from each other area, and one member from the IAB.
Diagrammatically, for an IETF with 3 areas, and fully filled areas:
Leadership Team
Area 1 Area 2 Area 3
Area Supervisor Area Supervisor Area Supervisor
Area 1 Area 1 Area 2 Area 2 Area 3 Area 3
Area Council Review Team Area Council Review Team Area Council Review team
Alice Charlie Charlie Alice Eric Bob
Bob Eric Dave Frank Frank Dave
IAB1 IAB2 IAB3
Terminology
-----------
Leadership - All the people involved in the process described here.
All the functions described here, except for those done by IAB members,
are
currently done by the IESG, so "IESG" is another possible name for this
group.
Leadership Team - The team of Area Supervisors and the IETF and IAB Chairs
Area Council - the people running an area
Review Teams - people doing document review. One team reviews each document.
Area Council: WG management
---------------------------
The area council is jointly responsible for the management of IETF work
related to the area it supervises. This includes all aspects currently
managed by ADs, except for final approval of working group creation,
charter change and shutdown, which rests with the technical leadership team.
Each WG is expected to have a single council member who is its primary
contact; normally, closely related WGs will have the same council member as
primary contact.
The area supervisor is expected to work chiefly on cross-WG matters.
Review team: Document approval
------------------------------
The review team is headed by the area supervisor, or (if desirable) a
review supervisor designated by the area supervisor.
It consists of one council member from each of the other areas. One council
member may, if desirable, serve on multiple review teams for other areas;
this may occur, for instance, where one area does not have enough activity
to require a large council.
The review team for an area is charged with doing cross-area final review
of documents, and ensure that documents conform to the published
requirements for the IETF publication form that working group and
standards-track documents are held to, as well as being useful for the
Internet.
If a review team has consensus on approving a document or returning it to
the WG, their decision is final (unless appealed); if a review team is
unable to reach consensus on a document, the document may be forwarded to
the technical leadership team, who can function as a backup review team for
the documents.
<UNCERTAIN> It may be valuable to add another area council member to the
review team, as a backup to the review supervisor. </UNCERTAIN>
The leadership team
-------------------
The leadership team, consisting of one area supervisor for each IETF area,
the IETF Chair and the IAB chair, is responsible for:
- Having an overall view of the strategy of the IETF in addressing the
challenges faced by the Internet where the IETF can provide useful input
- As part of maintaining that view, making final approval of WG charters,
changes to WG charters and WG termination
- Document clearly what the approval criteria of the IETF standards process
on documents are
- Serve as a "buck-stops-here" final arbitration function when disagreement
occurs within the IETF leadership about what the right thing to do on
documents is.
The executive function
----------------------
Other proposals for change have suggested creating an executive function to
deal with administrative issues of the IETF. This is not described in this
proposal, because the concerns are largely orthogonal to the structure
described here.
Details
-------
All members of the leadership are selected by the nomcom for membership in
an area.
The nomcom also selects which member is supervisor an area.
<UNCERTAIN>The supervisor may also be selected by the members of the
council, or by other means. Yearly selection by the council? If a
supervisor steps down, should the council or the leadership team be able to
select an interim supervisor </UNCERTAIN>
Special care should be taken that the composition of area teams and the
leadership team results in functional teams.
The composition of the review teams is decided by the leadership team.
The whole leadership meets at "leadership retreats", held once a year, and
once at every IETF meeting. Teleconferences in the review teams and area
teams are expected to happen roughly once a month; the leadership team
should meet every 2 weeks.
Discussion
----------
Among the important properties of the IETF is that the leadership is in
daily touch with the stuff being worked on, and that the final technical
approval rests in the hands of people with a wide range of perspectives,
all grounded in a common vision for the Internet.
This is something we have achieved today, by centralizing all process
oversight and final approval to the IESG, and something we do not want to
lose.
This reorganization attempts to preserve this by keeping document review
cross-area, and keeping the review in the hands of people who work actively
in bringing the IETF work forward. It attempts to disambiguate the role of
the AD (which is both being "WG advocate" and "process reviewer"), so that
it is very clear when the members act in one role or the other.
Finally, it attempts to reduce the overall load on individual IESG members;
the total number of hours spent on IETF work will increase under this
proposal, because all bodies need to spend significant time making sure
they work well together, but this is thought to be a price worthy of the
return.
Problems:
- The number of people we trust with making decisions grows by a rather
large amount.
- The training that happens today on the IESG is that people watch other
people do review, and learn a lot from that, including level-setting on the
difference between "important" problems and "unimportant" problems.
- The learning effect of having to review documents from many different
areas is substantial. If we review only docs from a single area, that's
lost. A suggestion to circulate members between areas might help that, but
also reduces consistency between review cycles when the membership of the
review team for an area changes.
- The issue of different review teams giving different feedback is
important. Consistency is not something we want to lose.
- If we improve the review this much, are we increasing people's tendency
to "leave the nit-finding to the review", or are we encouraging them to
"engineer to a known quality level"?