THE H-REPORT

A Q and A With Tony Rutkowski

by Christy Hudgins-Bonafield

Edited e-mail response from Internet Society (ISOC) executive director Tony Rutkowski to questions posed by Christy Hudgins-Bonafield.

Q: Internet Engineering Task Force (IETF) members are telling me that the potential exists for the ISOC to be untethered from IETF, that Christian Huitema is studying the relationship and that the main issue seems to be that ISOCÍs main purpose was to provide services, such as insurance in case of a lawsuit. The IETF members say that ISOC initially purchased the wrong kind of insurance, a policy that would not indemnify in case of a suit. Would you comment on this?

A: There is a lot of mis- and disinformation being propagated here. One could write a book on the intrigue and machinations going on in this area. To understand whatÍs occurring, some historical basics are needed.

Until a few years ago, the standards were developed in the IETF and passed on to the Internet Architecture Board (IAB), which consisted of long-standing Internet-community insiders. It played a kind of approval and appeal role. There was an IESG (IETF Steering Group) that consisted of the IETF area chairs. There was an IETF secretariat provided by Bob KahnÍs Corporation for National Research Initiatives (CNRI), funded as a U.S. government research project. In addition, a fair number of the people who played significant roles in the IETF were funded by various U.S. government sources.

The arrangement worked well, both for developing good new standards quickly an d in generating new applications and ideas. There were, however, several problems with this configuration.

1). There was no real organization. In legal terms, the IETF and the IAB are unincorporated associations; that is, they are people who simply meet from time to time. Even the secretariat is literally just a funded research project.

2). The standards had no international stature„a matter of considerable concern several years ago when the TCP/IP-versus-OSI battles were being waged in the United States and abroad. In addition, the lack of formal international stature gave rise to trade-barrier problems.

3). The U.S. government funds were likely to dry up eventually.

4). There was a remote possibility of individual antitrust liability if standards were adopted in a manner which ran afoul of established antitrust doctrine and the party affected by that subsequently undertook litigation.

One of the original purposes behind the creation of the ISOC was to have it serve as an international organizational umbrella for Internet standards infrastructure. Indeed, some trustees see this as the only significant purpose of ISOC, and corporate contributions to ISOC were sought based on [its] supporting the IETF.

For a variety of reasons, despite more than four years of trying, this has not occurred. The following chronicles some of the history:

In early 1992, the IAB agreed to have an association with ISOC as a body within ISOC. ISOC would give final approval to IAB members, and ISOCÍs Board would serve as an appellate body for its decisions.

ISOC began publishing adopted Internet standards and Requests for Comment in its membership publication and tried to establish a liaison relationship with international standards organizations on behalf of the IAB and IETF.

A process called POISED was launched to develop procedures for adopting standards and to deal with potential intellectual-property issues. But the results were perceived by many as being overly complicated and partially unnecessary.

In late 1992, the IAB unilaterally attempted to adopt an OSI-like solution to the IP next-generation problem. The resulting backlash from the IETF rank and file stripped the IAB of most functions and effectively placed the IESG in control of the IETF. A period of general distrust of and alienation from ISOC then ensued, with IETF members criticizing ISOC for acting from the top down, being a closed club of former insiders and attempting to take undeserved credit for the IETFÍs work. Bear in mind that the IETF regulars are very independent-spirited folks.

Attempts by ISOC to assume responsibility for the IETF secretariat were rejected by ISOC board member Bob Kahn, whose CNRI organization also holds the U.S. government contract, worth $75,000 annually, for providing the secretariat services. Instead, Kahn convinced the ISOC board to provide $225,000 to CNRI in 1993; CNRI, in turn, apparently gave the money to the U.S. government. This was done to allow ISOC to be able to say it was supporting the IETF.

In mid-1994, at an IETF meeting, someone apparently raised the issue of why there should be any ISOC connection; that person was told it was because ISOC provides them liability insurance. The statement was not accurate. The correct answer was that in any instances of anitrust litigation, the IETF would not be named as a defendant because it was a formal body of the ISOC; the ISOC would be designated the defendant rather than any individuals.

And so began an effort to get ISOCÍs insurance carrier to provide liability insurance coverage to those associated with the IAB/IESG/IETF. The carrier agreed to provide coverage under ISOCÍs existing liability insurance, up to a maximum of $2 million for ISOC and its trustees and officers, the IAB and IESG members, and ISOC itself. The carrier would not provide coverage to all of the IETF; the group was considered to be too amorphous since it was an unincorporated association.

In early 1995, ISOC agreed to fund separate legal counsel for the IETF, who in turn suggested that the ISO C-developed intellectual-property provisions be changed. In mid-1995, the ISOC board rejected a plan to hire a staff member to assist the IETF/IESG, but voted to provide to IETF/IESG $125,000 in 1995 and $250,000 in subsequent years, with no stipulations for the money. This was done to demonstrate interest in supporting the IETF.

[Christian] Huitema (former IAB chairperson and now an ISOC board member) is supposed to be developing a plan for bringing the IETF under ISOC. Meanwhile, the IETF rank and file has expressed, both publicly and privately, its general disinterest in, if not antipathy toward, ISOC. Discussion began within the IETF about incorporating separately and making the IETF secretariat self-funded instead of a U.S. government-funded research project under CNRI.

One final point: The Internet standards development scene has changed significantly. The OSI-versus-TCP/IP question or even the formal international standards problems are largely nonexistent. The emerging World Wide Web Consortium has taken on some of the most important new Internet standards activities. Increasingly, vendors themselves are creating fairly broad-based ad hoc standards groups to develop specific standards suites.

Q: What do you think is the likelihood that ISOC and IETF will split?

A: As noted, they were never joined! The question is which way things will proceed„further apart or further together. Based on my analysis during the past year, the former seems more likely. As both a trustee and the executive director of ISOC, IÍve argued during the years for an alliance approach rather than ownership and control, an approach that would provide real assistance and help IETF to be independent if thatÍs what the membership wants. This, however, was a minority view on the board.

Q: Did ISOC indeed purchase the wrong kind of insurance? If so, has this been remedied?

A: This was not the case. The only question was about the extent of the groups covered. This was handled through a policy rider, in March 1993, that explicit ly named the IAB and IESG.

Q: When is Christian HuitemaÍs report due? Do you have specific expectations?

A: Christian represents a group that earnestly hopes for a significant relationship under ISOC. In addition to the factors noted above, I think there are some significant differences in age and perspective. Some have compared [the differences] to elderly parents dealing with teenage children.

Q: Do you know of factions within IETF that believe certain tasks now under ISOC should be handled elsewhere? Things like finding funding for IETF outside of the U.S. government, setting up procedures that will make IETF activities less prone to lawsuits, overseeing international coordination with standards bodies or determining the effect of IETF activity on procurement policy?

A: As noted, there isnÍt really much under ISOC. The real big-ticket item is the IETF secretariat under CNRI and the U.S. government.

YouÍll likely find most people on that side, with various shades and degrees of vehemence! Carl Malamud headed the original POISED process. Steve Knowles of FTP Software has been adamantly against any ISOC role. The current IETF chairperson, Paul Mockapetris, has been generally hostile„which is particularly interesting considering heÍs one of the original DARPA Internet insiders. SameÍs true for Jon Postel.

ISP legend Rick Adams, who is both a board member of the CIX Association and an officer of ISOCÍs advisory board and is allowed to sit in on open board meetings, can offer good perspective. Rick is one of the few high-level people in the industry who is technically and operationally competent and usually tells it like it is. If anything, there may be an argument today that the IETF should have an affiliation with the CIX Association, given the fact that the standards most significantly affect the ISPs.

SprintÍs Sean Doran is an example of a new, outspoken group of really good Internet engineers who have disdain for the ñold boys.î HeÍs been outspoken on com-priv. I thought his recent note rather cl ever where he came up with ñISOC=It Seeks Overall Control.î

October 15, 1995