Network Working Group T. Hardie
Request for Comments: 3929 Qualcomm, Inc.
Category: Experimental October 2004
Alternative Decision Making Processes
for Consensus-Blocked Decisions in the IETF
Status of this Memo
This memo defines an Experimental Protocol for the Internet
community. It does not specify an Internet standard of any kind.
Discussion and suggestions for improvement are requested.
Distribution of this memo is unlimited.
Copyright Notice
Copyright (C) The Internet Society (2004).
Abstract
This document proposes an experimental set of alternative decision-
making processes for use in IETF working groups. There are a small
number of cases in IETF working groups in which the group has come to
consensus that a particular decision must be made but cannot agree on
the decision itself. This document describes alternative mechanisms
for reaching a decision in those cases. This is not meant to provide
an exhaustive list, but to provide a known set of tools that can be
used when needed.
1. Introduction
Dave Clark's much-quoted credo for the IETF describes "rough
consensus and running code" as the key criteria for decision making
in the IETF. Aside from a pleasing alliteration, these two
touchstones provide a concise summary of the ideals that guide the
IETF's decision making. The first implies an open process in which
any technical opinion will be heard and any participant's concerns
addressed; the second implies a recognition that any decision must be
grounded in solid engineering and the known characteristics of the
network and its uses. The aim of the IETF is to make the best
possible engineering choices and protocol standards for the Internet
as a whole, and these two principles guide it in making its choices
and standards.
In a small number of cases, working groups within the IETF cannot
reach consensus on a technical decision that must be made in order to
ensure that an interoperable mechanism or set of standards is
Hardie Experimental [Page 1]
RFC 3929 Consensus-Blocked Decisions in the IETF October 2004
available in some sphere. In most of these cases, there are two or
more competing proposals at approximately the same level of technical
maturity, deployment, and specification. In some cases, working
groups can achieve consensus to advance multiple proposals and either
to revisit the question with experience or to build the required
mechanisms to handle multiple options for the life of the protocol.
In other cases, however, a working group decides that it must advance
a single proposal.
Choosing among proposals can be difficult especially when each is
optimized for slightly different use cases, as this implies that the
working group's best choice depends on the participants' views of
likely future use. Further problems arise when different proposals
assign costs in implementation, deployment, or use to different
groups, as it is a normal human reaction to seek to prevent one's own
ox from being gored.
This document proposes a set of experimental mechanisms for use in
such cases. To gauge the results of the use of these mechanisms, the
Last Call issued to the IETF community should note such a mechanism
is being used and which proposal among the set was chosen. If and
when the community becomes satisfied that one or more of these
methods is useful, it should be documented in a BCP.
In no way should this experiment or any future BCP for this small
number of cases take precedence over the IETF's normal mode of
operation.
2. Rough Consensus as a baseline approach
The Conflict Research Consortium at the University of Colorado
outlines the pros and cons of consensus as follows:
The advantage of consensus processes is that the resulting
decision is one that meets the interests of all the parties and
that everyone can support. The disadvantage is that developing
such a decision can be a very slow process, involving many people
over a long period of time. There is also a relatively high
probability of failure. If a quick decision is needed, the
consensus approach may not work. Consensus rule processes also
tend to favor those that oppose change and want to preserve the
status quo. All these people have to do is refuse to support any
consensus compromises and they will win (at least as long as they
can delay change) [CONFLICT].
Hardie Experimental [Page 2]
RFC 3929 Consensus-Blocked Decisions in the IETF October 2004
Using "rough consensus" as a guideline limits some of the
disadvantages of consensus processes by ensuring that individuals or
small factions cannot easily block a decision that otherwise has
general support. The touchstone of "running code" can also limit the
disadvantages of consensus processes by requiring that statements
opposing particular proposals be technically grounded.
These limitations do not change the core mechanisms of consensus-
building, however, and the IETF process continues to require
individual participants both to use their best engineering judgment
to select among proposals and to balance their own interests with
those of the Internet as a whole. Active participation and a
willingness to compromise, possibly on key points, are needed.
Historically, this has worked because a large majority of
participants have recognized that the Internet's growth and
enhancement are more important overall than any specific short-term
advantage.
In other words, "rough consensus" is sufficient in most cases in the
IETF to ensure not only that individuals or small groups are heard
when they raise technical objections, but also that they cannot block
progress when general agreement has been reached. This document does
not suggest changing the usual mechanisms for achieving progress; it
proposes mechanisms for use when a working group has consensus that
it must make a decision but cannot make that decision by the usual
rules.
3. Conditions for use
In general, working groups should consider using alternate decision-
making processes when it is clear both that a choice must be made and
that the choice cannot be made with continued discussion, refinement
of specifications, and implementation experience. A guideline for
determining whether these conditions have been met is included below.
3.1. There is a clear decision to be reached
There must be a clear statement of the decision to be reached. This
may be in the working group's charter, in requirements documents, or
in other documents developed by the working group. Prior to any
invocation of an alternate decision making process, the Chair(s)
should confirm with the working group that there is general agreement
on the decision to be reached. This should include a specific
consensus call on whether the working group can advance multiple
proposals or must select a single proposal for the work item.
Hardie Experimental [Page 3]
RFC 3929 Consensus-Blocked Decisions in the IETF October 2004
3.2. Proposals are available in Draft form
Proposed solutions must be available as Internet-Drafts and must be
sufficiently specified so that the Chair(s) believe they could be
published as an IETF specification, possibly with further refinement.
If the Chair indicates that a proposed solution is insufficiently
specified, concrete problems must be identified, and a reasonable
amount of time provided to resolve those problems must be provided.
Note that if one of the proposed solutions is "do nothing", an
explicit Draft to that effect must be available; it may, however, be
produced when the group invokes an alternate decision-making process.
3.3. The working group has discussed the issue without reaching
resolution
Consensus-building requires significant amounts of discussion, and
there is no general rule for indicating how much discussion a
technical issue requires before a group should reach consensus. If
there is any question about whether the discussion has been
sufficient, the working group chair(s) should always err on the side
of allowing discussion to continue. Before using an alternate
decision making process, the working group chair(s) should also make
an explicit call for consensus, summarizing the technical issues and
the choice to be made. If new technical points are made during the
call for consensus, discussion should continue. If no new points are
raised, but the group cannot come to consensus, the working group may
consider using an alternate decision making process. Under no
circumstances is the working group required to use an alternate
decision-making process.
3.4. There is an explicit working group last call to use an alternate
method
In item 3.3 above, it is noted that the Chair(s) should make an
explicit call for consensus on the technical issues and should
proceed only after that call has yielded no forward progress. A
different Last Call on whether to use an alternate decision-making
method is required, with a stated period for comments by working
group members. This is to indicate that the decision to use an
alternate method should be taken at least as seriously as the
decision to advance a document on the standards track. It also
provides a clear signal that this is a last moment for participants
to reconsider their positions. The decision to use an alternate
decision making process requires the rough consensus of the working
group, as determined by the Chair(s). The choice of which process to
use may be made in the Last Call or may be the subject of separate
discussions within the working group. If the group comes to
Hardie Experimental [Page 4]
RFC 3929 Consensus-Blocked Decisions in the IETF October 2004
consensus that an alternative method is required but does not come to
consensus on the method to use, an external review team (c.f. section
4.1, below) will be formed.
In discussions regarding this document, several points have been
raised about the viability of any mechanism that requires consensus
to use an alternative to consensus-based decision making. Some
individuals have pointed out that groups having trouble achieving
consensus on a technical matter may have similar problems achieving
consensus on a procedural matter. Others have been concerned that
this will be used as an attempt to end-run around rough consensus.
These are valid concerns, and they point both to the need to retain
rough consensus as the baseline mechanism and the need to exercise
caution when using these alternate methods. More importantly though,
they highlight the nature of these alternatives. They are primarily
mechanisms that allow people to recognize the need for compromise in
a new way, by backing away from entrenched technical positions and by
putting the technical choice in the hands of the broader community.
They highlight that the choice for each participant is now between
achieving a result and failure.
There is a fundamental tension between the IETF community's desire to
get the best decision for a particular technical problem and its
desire to get a decision that has community buy-in in the form of
rough consensus. These mechanisms cannot resolve that fundamental
tension. They may, however, provide a way forward in some situations
that might otherwise end in a deadlock or stagnation.
4. Alternate Methods
In setting up an alternate method, care must be taken that the
process by which the decision is reached remains open and focused on
the best technical choice for the Internet as a whole. The steps set
out below provide a straw proposal for four such mechanisms. These
systems are relatively heavyweight, partially to highlight the
gravity of invoking these methods and partially to ensure that the
IETF community as a whole is alerted to and kept informed of the
process. Note that alternate procedures have been used in the past;
see [RFC3127] for a description of that used in the decision between
two competing candidate protocols for Authentication, Authorization,
and Accounting. By setting out these proposals, this document does
not intend to limit working group choice but intends to provide a set
of well-defined processes that obviate the need for reinvention in
most cases.
Hardie Experimental [Page 5]
RFC 3929 Consensus-Blocked Decisions in the IETF October 2004
4.1. Alternate Method One: External Review Team Formation
The working group notifies the IETF community that it intends to form
an external review team by making a public announcement on the IETF-
announce mailing list. That announcement should include a summary of
the issue to be decided and a list of the Internet-Drafts containing
the alternate proposals. It should also include the name and
location of an archived mailing list for the external review team's
deliberations.
4.1.1. External Review Team Membership
External review teams have five members who must meet the same
eligibility requirements as those set out for a voting member of the
NomCom [RFC3777]. Explicitly excluded from participation in external
review teams are all those who have contributed to the relevant
working group mailing list within the previous twelve months, the
IESG, the IAB, and the members of an active NomCom.
Volunteers to serve on the review team send their names to the IETF
executive director. Should more than five volunteer, five are
selected according to the process outlined in [RFC3797]. Note that
the same rules on affiliation apply here as to the NomCom, to reduce
the burden on any one organization and to remove any implication of
"packing" the review team.
Participants in the working group may actively solicit others to
volunteer to serve on the review team but, as noted above, they may
not serve themselves if they have commented on the list within the
previous twelve months.
4.1.2. External Review Team Deliberation
The external review team is alloted one month for deliberations. Any
member of the team may extend this allotment by two weeks by
notifying the relevant working group Chair(s) that the extension will
be required.
The team commits to reading the summary provided during the IETF
announcement and all of the relevant Internet-Drafts. Members may
also read the archived mailing list of the working group and may
solicit clarifications from the document authors, the working group
chairs, or any other technical experts they choose. All such
solicitations and all deliberations among the review team of the
proposals should take place on the archived mailing list mentioned in
the IETF announcement. The team members may, of course, have one-
on-one discussions with relevant individuals by phone, email, or in
person, but group deliberations should be on the archived list.
Hardie Experimental [Page 6]
RFC 3929 Consensus-Blocked Decisions in the IETF October 2004
4.1.3. Decision Statements
Each member of the external review team writes a short decision
statement, limited to one page. That decision statement contains a
list of the proposals in preference order. It may also contain a
summary of the review team member's analysis of the problem and
proposed solutions, but this is not required. These decision
statements are sent to the archived mailing list, the relevant
working group chair(s), and the IESG.
4.1.4. Decision Statement Processing
The decision statements will be tallied according to "instant-runoff
voting" rules, also known as "preference voting" rules [VOTE].
4.2. Alternate Method Two: Mixed Review Team
This mechanism allows the working group to designate a review team
that involves those outside the working group and those who have been
involved in the process within the working group. Although it may
appear that having a single representative of each proposal will have
a null effect on the outcome, this is unlikely, except when there is
a binary choice, because of the rules for decision-statement
processing (c.f. 4.1.4.). As in 4.1, the working group notifies the
IETF community that it intends to form a mixed review team by making
a public announcement on the IETF-announce mailing list. That
announcement should include a summary of the issue to be decided and
a list of the Internet-Drafts containing the alternate proposals. It
should also include the name and location of an archived mailing list
for the external review team's deliberations.
4.2.1. Mixed Review Team Membership
Mixed review teams are composed of one designated representative of
each of the proposals, typically the Internet-Draft's principal
author, and six external members. Five of the external members are
selected per 4.1.1. above. The sixth is designated by the IESG as a
chair of the group. Though the primary role of the chair is to
ensure that the process is followed, she or he may vote and engage in
the deliberations.
4.2.2. Mixed Review Team Deliberation
The review team is alloted one month for its deliberations, and any
member of the team may extend that allotment by two weeks by
notifying the review team Chair this the extension will be required.
Hardie Experimental [Page 7]
RFC 3929 Consensus-Blocked Decisions in the IETF October 2004
The review team commits to reading the summary provided during the
IETF announcement and all of the relevant Internet-Drafts. Members
may also read the archived mailing list of the working group, and of
any other technical experts as they see fit. All such solicitations
and all deliberations among the review team of the proposals should
take place on the archived mailing list mentioned in the IETF
announcement.
4.2.3. Decision Statements
As in 4.1.3, above.
4.2.4. Decision Statement Processing
As in 4.1.4, above.
4.3. Alternate Method Three: Qualified Short-Straw Selection
As in 4.1 and 4.2, the working group notifies the IETF community that
it plans to use an alternate decision mechanism by making a public
announcement on the IETF-announce mailing list. That announcement
should include a summary of the issue to be decided and a list of the
Internet-Drafts containing the alternate proposals.
In this method, a single working group participant is selected to
make the decision. Any individual who has contributed to the working
group in the twelve months prior to the working group Last Call on
the technical question (c.f. 3.3, above) may volunteer to serve as
the decision maker. Individuals may qualify as participants by
having made a public statement on the working group mailing list, by
serving as an author for an Internet-Draft under consideration by the
working group, or by making a minuted comment in a public meeting of
the working group. The Chair(s) may not volunteer. Each qualified
volunteer sends her or his name to the working group chair and the
IETF Executive Director within three weeks of the announcement sent
to the IETF-announce mailing list. The IETF Executive Director then
uses the selection procedures described in [RFC3797] to select a
single volunteer from the list. That volunteer decides the issue by
naming the Internet-Draft containing the selected proposal in an
email to the relevant working group chair, the working mailing list,
and the IESG.
4.4. Alternate Method Four: Random Assignment
Among the small number of cases for which consensus is not an
appropriate method of decision-making are an even smaller number for
which the decision involves no technical points at all and a need to
select among options randomly. The IDN working group, as an example,
Hardie Experimental [Page 8]
RFC 3929 Consensus-Blocked Decisions in the IETF October 2004
needed to designate a specific DNS prefix. As the decision involved
early access to a scarce resource, a random selection was required.
In such cases, a working group may ask IANA to make a random
assignment from among a set of clearly delineated values. Under such
circumstances, IANA will be guided by [RFC3797] in its selection
procedures. Under extraordinary circumstances, the working group
may, with the approval of the IESG, ask IANA to select among a pool
of Internet-Drafts in this way.
5. Appeals
The technical decisions made by these processes may be appealed
according to the same rules as any other working group decision, with
the explicit caveat that the working group's consensus to use an
alternate method stands in for the working group's consensus on the
technical issue.
6. Security Considerations
The risk in moving to a system such as this is that it shifts the
basis of decision making within the IETF. In providing these
mechanisms, it is hoped that certain decisions that may be
intractable under consensus rules may be reached under the rules set
out here. The risk, of course, is that forcing the evaluation to
occur under these rules may allow individuals to game the system.
7. IANA Considerations
Section 4.3 may require the IANA to make random selections among a
known set of alternates.
8. References
8.1. Normative References
[RFC3797] Eastlake, D., "Publicly Verifiable Nomination Committee
(NomCom) Random Selection", RFC 3797, June 2004.
[RFC3777] Galvin, J., Ed. "IAB and IESG Selection, Confirmation, and
Recall Process: Operation of the Nominating and Recall
Committees", BCP 10, RFC 3777, June 2004.
8.2. Informative References
[RFC3127] Mitton, D., StJohns, M., Barkley, S., Nelson, D., Patil,
B., Stevens, M., and B. Wolff, "Authentication,
Authorization, and Accounting: Protocol Evaluation", RFC
3127, June 2001.
Hardie Experimental [Page 9]
RFC 3929 Consensus-Blocked Decisions in the IETF October 2004
[VOTE] Center for Democracy and Voting. "Frequently Asked
Questions about IRV", http://www.fairvote.org/irv/faq.htm.
[CONFLICT] International Online Training Program on Intractable
Conflict,"Consensus Rule Processes", Conflict Research
Consortium, University of Colorado, USA.
http://www.colorado.edu/conflict/peace/treatment/
consenpr.htm
10. Acknowledgements
The author would like to acknowledge the contributions and
challenging exchanges of those who reviewed this document, among them
John Klensin, Dave Crocker, Pete Resnick, Spencer Dawkins, Scott
Bradner, Joel Halpern, Avri Dora, Melinda Shore, Harald Alvestrand,
Alex Simonelis, Keith Moore, Brian Carpenter, and Alex Rousskov.
11. Author's Address
Ted Hardie
Qualcomm, Inc.
675 Campbell Technology Parkway
Suite 200
Campbell, CA U.S.A.
EMail: hardie@qualcomm.com
Hardie Experimental [Page 10]
RFC 3929 Consensus-Blocked Decisions in the IETF October 2004
Full Copyright Statement
Copyright (C) The Internet Society (2004).
This document is subject to the rights, licenses and restrictions
contained in BCP 78, and except as set forth therein, the authors
retain all their rights.
This document and the information contained herein are provided on an
"AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET
ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE
INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED
WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Intellectual Property
The IETF takes no position regarding the validity or scope of any
Intellectual Property Rights or other rights that might be claimed to
pertain to the implementation or use of the technology described in
this document or the extent to which any license under such rights
might or might not be available; nor does it represent that it has
made any independent effort to identify any such rights. Information
on the IETF's procedures with respect to rights in IETF Documents can
be found in BCP 78 and BCP 79.
Copies of IPR disclosures made to the IETF Secretariat and any
assurances of licenses to be made available, or the result of an
attempt made to obtain a general license or permission for the use of
such proprietary rights by implementers or users of this
specification can be obtained from the IETF on-line IPR repository at
http://www.ietf.org/ipr.
The IETF invites any interested party to bring to its attention any
copyrights, patents or patent applications, or other proprietary
rights that may cover technology that may be required to implement
this standard. Please address the information to the IETF at ietf-
ipr@ietf.org.
Acknowledgement
Funding for the RFC Editor function is currently provided by the
Internet Society.
Hardie Experimental [Page 11]
RFC, FYI, BCP