| Internet-Draft | RFC8989bis | December 2022 |
| Duke | Expires 4 June 2023 | [Page] |
The IETF Nominating Committee (NomCom) appoints candidates to most IETF leadership committee. RFC8713 provides criteria for membership on NomCom that attempt to ensure that NomCom volunteers are members of the loosely defined IETF community, by requiring in-person attendance in three of the past five in- person meetings. In 2020 and 2021, the IETF had six consecutive fully online plenary meetings that drove rapid advancement in remote meeting technologies and procedures, including an experiment that included remote attendance for NomCom eligibility. This document updates RFC8713 by building a new set of eligibility criteria from first principles, with consideration for the increased salience of remote attendance.¶
This note is to be removed before publishing as an RFC.¶
Source for this draft and an issue tracker can be found at https://github.com/ietf-wg-elegy/rfc8989bis.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶
This Internet-Draft will expire on 4 June 2023.¶
Copyright (c) 2022 IETF Trust and the persons identified as the document authors. All rights reserved.¶
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
[RFC8713] defines the process for selection of the Internet Architecture Board (IAB), Internet Engineering Steering Group (IESG), IETF Trust, and one IETF LLC Director. A key actor in the process is the Nominating Committee (NomCom), which nominates a single candidate for each open position, subject to confirmation by other bodies.¶
NomCom voting members are volunteers that have met certain eligibility requirements. The actual NomCom is selected at random from the pool of eligible volunteers. Thus, it is important that members of the pool be IETF participants likely to have knowledge of IETF processes and Tao. There are restrictions to ensure that no more than two volunteers with the same primary affiliation are chosen.¶
Section 4.14 of [RFC8713] requires that volunteers must have attended three of the previous five meetings. In practice, this has meant that the volunteer picked up their registration badge. Current members of the Internet Society Board of Trustees and bodies for which the NomCom nominates members are ineligible.¶
[RFC8989] specified an experiment in the wake of six consecutive fully online meetings from 2020 to 2021, where the traditional interpretation of the requirement would have resulted in no eligible volunteers. It extended the attendance requirement to define meeting attendance as including logging in to at least one session of a fully-online IETF meeting.¶
RFC8989 also created two other tracks to obtain eligibility: (1) serving as a working group chair or secretary in the past 3 years, and (2) author or editor of an IETF Stream RFC in the past five years, including internet-drafts in the RFC Editor queue.¶
This document discusses some of the first principles that inform the design of NomCom eligibility, and makes recommendations on how the process of qualification based on attendance should work.¶
This document replaces the attendance criteria in Section 4.14 of [RFC8713] with criteria based on those in [RFC8989], and obsoletes RFC8989 to make it clear that that document has been superceded. The other paragraphs of Section 4.14 of [RFC8713] remain intact.¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
The NomCom is intended to be composed of randomly selected members of "the community." For many years, in-person attendance was a reasonable proxy for the commitment associated with being a member. Two days of travel and an attendance fee is a relatively large expenditure of time and money. Additionally, in-person attendance is thought to increase personal familiarity with candidates for leadership positions and with the spirit of the IETF, although there is no mechanism to ensure any interactions. Finally, the NomCom interview process was largely conducted in-person at IETF meetings, so the ability to attend was a prerequisite to participate.¶
Beyond the principle that the community should govern itself, selecting volunteers with a demonstrated commitment to the organization, while limiting the number from any organization, avoids the potential for mischief via nominations that disrupt IETF operations or work against the interests of the community as a whole.¶
However, attitudes to business travel evolve, and remote meeting technology continues to improve, to the extent that many longstanding community members choose to participate remotely. The system has always excluded some from attendance, and thus qualification from the NomCom, due to cost or personal reasons. Further, the NomCom can now fully complete its business using online tools.¶
Counting remote attendance lowers the barriers to entry. As IETF is committed to having a no-fee remote option ([I-D.ietf-shmoo-remote-fee]) the only required investment is to log on once per meeting at a specific time (sometimes a locally inconvenient hour). While this document does not formally impose a requirement for the NomCom to function entirely remotely, including remote-only attendees in the pool is likely to effectively require a remote component to NomCom operations.¶
Finally, overly restrictive criteria work against getting a deep talent pool.¶
The following paths to qualification replace Section 4.14 of [RFC8713]. Any one of the paths is sufficient, unless the person is otherwise disqualified under Section 4.15 of [RFC8713].¶
Path 1: The person has registered for and attended 3 out of the last 5 IETF meetings, either in-person or online. In-person attendance is as determined by the record keeping of the Secretariat. Online attendance is based on being a registered person who logged in for at least one session of an IETF meeting.¶
Path 2: The person has been a Working Group Chair or Secretary within the 3 years prior to the day the call for NomCom volunteers is sent to the community.¶
Path 3: The person has been a listed author or editor (on the front page) of at least two IETF Stream RFCs within the last 5 years prior to the day the call for NomCom volunteers is sent to the community. An Internet-Draft that has been approved by the IESG and is in the RFC Editor queue counts the same as a published RFC, with the relevant date being the date the draft was added to the RFC Editor queue. For avoidance of doubt, the 5-year timer extends back to the date 5 years before the date when the call for NomCom volunteers is sent to the community.¶
The most potent threat associated with NomCom eligibility is that an organization or group of coordinating organizations would attempt to obtain a majority of NomCom positions, in order to select an IETF leadership in support of an agenda that might be self-serving and against the interests of the community as a whole.¶
Note that [RFC8713] lets the Chair decide the NomCom voting requirement, so a simple majority may be inadequate. However, 7 of 10 forms a quorum, so at worst seven NomCom members working together can almost certainly impose their will.¶
Whatever the merits of admitting remote attendees, it reduces the minimum cost of creating a NomCom-eligible volunteer from three in-person trips of ~5 days each over the course of at least 8 months, to zero financial cost and the time required to log in three times over at least 8 months. Some organizations might not be deterred in either case, while others might now find such an attack to not be feasible.¶
A large number of "legitimate" volunteers makes it quite difficult to control 6 of 10 NomCom slots. Setting aside limitations on the number of selections from any organization, basic probability shows that to have even a 50% chance of controlling 6 or more NomCom positions, an attacker needs somewhat roughly 60% of the volunteer pool. For example, if there are 300 "legitimate" volunteers, an attacker must produce 365 volunteers to exceed a 50% chance of NomCom capture (see Appendix A).¶
A sudden surge in the number of volunteers, particularly of people that no one recognizes as a part of the community, is an early-warning sign for the community, leadership and the IETF Secretariat to further investigate. The IETF should monitor and assess a sudden increase in the number of online registration fee waivers awarded in accordance with Section 4 of [I-D.ietf-shmoo-remote-fee].¶
While loosening eligibility criteria lowers the cost to an attacker of producing eligible volunteers, it also increases the number of "legitimate" volunteers that increases the difficulty of an attack.¶
The two-per-organization limit in [RFC8713] complicates such an attack. To circumvent it, an organization must either (1) coordinate with at least two like-minded organizations to produce a NomCom majority, (2) incentivize members of other organizations (possibly through a funding agreement) to support its agenda, or (3) propose candidates with false affiliations.¶
While the IETF does not routinely confirm the affiliation of volunteers, as part of an investigation it could eliminate volunteers who have misrepresented said affiliation. Publishing the list of volunteers and affiliations also gives the community an opportunity to review the truth of such claims.¶
Assuming that 300 legitimate volunteers are all from different organizations, three conspiring organizations would need 771 volunteers (257 per organization) for a 50% chance of NomCom capture (see Appendix A).¶
Attendance at 3 meetings requires at least 8 months. Given the volume of volunteers necessary to capture the process, an attack requires a surge in attendees over the course of a year. Such a surge might trigger a community challenge to the list of eligible volunteers, and/or a leadership investigation to detect suspicious behavior (e.g. logging in to a single session and then immediately logging out). In the event of abuse of process, the leadership would then have months to adjust policy in response before the NomCom cycle begins.¶
Note that the normalization of consistent remote participation allows for a single individual to mount an attack that previously required coordination. By registering for IETF meetings using a number of different identities over a year, an individual can make each of those identities NomCom eligible and then serve under any one of them that is selected for the NomCom. Once selected, an individual could seek to disrupt the process or prevent the timely conclusion of its work.¶
This attack is much harder to detect or prevent than equivalent attacks were previously, as it does not require coordination among multiple attendees. While the attacker cannot be sure of fee waivers for some or all of the different identities, the lower cost for remote participation also makes this attack more feasible than it would have been under prior rules.¶
However, the voting member recall procedure in Section 5.7 of [RFC8713] exists to allow removal and replacement of disruptive figures.¶
Additional changes to the process to further obstruct attacks against the NomCom are beyond the scope of this document. However, a challenge process against volunteers with a suspicious reported affiliation, or that might be aliases of a single volunteer, could trigger an investigation.¶
Similarly, the challenge to the random selection described in Section 4.17 of [RFC8713] can explicitly include appeals against the data used to qualify the volunteer, rather than the randomization process.¶
This document has no IANA actions.¶
Section 5 offers some mathematical results for the probability of NomCom capture. This appendix shows the work.¶
Let (a ch b) mean the number of combinations of b items chosen from a population of a items, or¶
(a ch b) = fact(a) / (fact(a-b) * fact(b))¶
The first computation assumes there is no limit of two per organization, or equivalently, no organization produces more than two volunteers.¶
Let L be the number of "legitimate" volunteers (i.e. those not allied with an attacker" and A be the number of attacking volunteers. Then there are ((L+A) ch 10) ways to select a NomCom. The number of outcomes where attackers capture the NomCom is¶
Sum(i=6..10)[(A ch i) * (L ch (10-i))]¶
and the probability of capture is therefore¶
Sum(i=6..10)[(A ch i) * (L ch (10-i))] / ((L+A) ch 10).¶
For L = 300, this probability crosses 50% at A = 365.¶
Assume that the population of L is drawn from L different organizations (this assumption is unfavorable to the attacker). Assume also that there are three conspiring organizations. Then no more than 6 members can be drawn from A.¶
Let B be the number of nominees per attacking organization, so that A = 3B.¶
The number of combinations to pick exactly N attackers, N <= 6, is¶
C(N) = (L ch (10-N)) * Sum(i=0:min(N,2))[(B ch i)Sum(j=0..min(2, N-i))[(B ch j)(B ch min(2, N-i-j))]]¶
And the probability of capture is¶
C(6) / Sum(i=0..6)[C(i)]¶
For L = 300, the A required to exceed a 50% probability of capture is 771.¶
RFC Editor's Note: Please remove this section prior to publication of a final version of this document.¶
Brian Carpenter and Stephen Farrell wrote RFC8989, which provides the core of this document.¶
Luc Andre Burdet, Brian Carpenter, and Donald Eastlake provided useful editorial suggestions.¶