Note: This is not a complete report of the plenary sessions; rather, it is a summary of the highlights of the discussions. All IETF 74 presentations can be found here.
The IETF Secretariat has been working on a redesign of the IETF Web site, which turned out to be a much bigger job than anticipated. IETF chair Russ Housley thanked the secretariat staff for their hard work.
He also reported on the Code Sprint from early in the week of the IETF meeting, which was a big success. Five releases were developed, including a new version of the datatracker, which improved the authorisation system and which replaced a number of hard-to-maintain legacy scripts.
Finally, Russ thanked all of the sponsors and contributors who had made the meeting possible.
IAOC chair Jonne Soininen and IETF administrative director Ray Pelletier updated participants on issues related to the financial and organisational status of the IETF. In their report, they stated that nearly 50 percent of the IETF’s revenues are derived from meeting registration fees, while 14 percent is met by contributions from meeting sponsors and Network Operations Centre (NOC) sponsors, which are secured by ISOC. ISOC provides nearly one-third of the annual funding necessary for IETF operations.
Less than half of the IETF’s expenses are allocated for meetings. Day-to-day secretariat expenses-including IT support, RFC Services, and Tools and Administrative costs-total more than USD 2.5 million annually, not including volunteer time and with no direct source of funding.
In 2008, expenses for most activities were under budget. However, while the meeting revenue for IETF 71 and IETF 72 were on target, meeting attendance for IETF 73 was relatively low. At USD 616,000, both IETF host and NOC sponsorships reached a record level.
In light of the current global economic uncertainties, the IAOC has developed financial contingency plans based on 15 percent and 25 percent attendance attritions, which will be evaluated throughout the course of the year. The Internet Society committed to provide a safety net for 2009 should there be an attendee shortfall. This has made it possible for the IAOC to keep the registration fee for IETF 74 at the current rate. Registration fees for future meetings will be reviewed during the year. The IAOC is looking into other opportunities to increase participation at future IETF meetings.
The difficulties associated with visa requirements for entering the United States, especially for people from China, are still concerns. The IAOC requested assistance from the U.S. State Department, without much success. One solution might be to reduce the number of meetings held in the United States, perhaps moving some to Canada instead. The topic was further discussed during the open-mic portion of the plenary, in which a number of people suggested further improvements to remote-participation facilities. Thomas Narten suggested setting up a design team, which would codify rules and behaviour expectations so as to make remote participation easier and more appealing. A mailing list intended to further work on this topic was created following the meeting and can be found here.
Ed Juscevicius gave his final report as chair of the IETF Trust prior to turning over the chair position to Marshall Eubanks. The Trust has done work on the legal provisions related to IETF documents. There is also a new frequently-asked-questions document on copyright issues. It is available here.
Patents at the IETF
IETF attorney Jorge Contreras and Scott Bradner gave a presentation describing a number of issues related to patents and disclosure obligations. While the rules can be found in RFC 3979, Scott and Jorge highlighted some of the relevant points with regard to the IETF’s patent policies. For example, in terms of disclosure obligations, they explained that an IETF participant must disclose any known patent that the participant (or the participant’s sponsor) controls and that may cover any IETF contribution. However, an IETF participant or anyone else may disclose third-party patents the person believes may cover IETF contributions.
Scott and Jorge also said that because IETF participation is by individuals, disclosure is primarily an individual obligation. Individual engineers must ensure that their employer companies make the required disclosures. If an engineer cannot ensure a disclosure, that engineer should not participate. Similarly, companies that own the patents may be deemed to control the actions of their participating employees.
Disclosure is required “as soon as reasonably possible”? after a contribution is published as an Internet-Draft. If this is your own contribution, disclosure should follow relatively quickly, such as within a few days. But Jorge and Scott suggest that you not make the contribution until you’re ready to file the disclosure. In addition, if an IETF participant first learns of a patent after publication of the affected Internet-Draft, a disclosure must be made as soon as reasonably possible after the discovery.
Disclosures may be updated voluntarily at any time, but they must be updated when an IETF document changes such that its coverage by a disclosed patent or patent application changes or if the claims of a patent or patent application are amended so that their coverage of IETF documents changes. Failure to comply with patent disclosure requirements is a violation of IETF policy, and the potential legal consequences to companies are considerable.
Both Jorge and Scott strongly recommended that IETF participants refer to RFC 3979 and consult their legal counsels.
Internet Architecture Board chair Olaf Kolkman updated participants on recent IAB activities, including the IAB having finalized two documents since the last meeting: Principles of Internet Host Configuration and Design Choices When Expanding DNS.
Other documents that have been worked on but are still under review include RFC Streams Headers and Boilerplates, the RFC Editor Model, IAB Thoughts on IPv6 Network Address Translation, P2P [Peer-to-Peer] Architectures, Defining the Role and Function of IETF Protocol Parameter Registry Operators, and Evolution of the IP Model.
The IAB continues to follow developments with respect to multiprotocol label switching-transport profile (MPLS-TP). At IETF 74, concerns were expressed about confusion in the marketplace regarding the utility and standardization status of transport multiprotocol label switching (T-MPLS). There have been claims that MPLS-TP is a forward-compatible update from T-MPLS. “This is clearly not the case,” said Olaf. The IAB is in the process of reviewing some of the liaison relations between the IETF and other organizations.
At the end of his IAB update Olaf announced the new IAB members and thanked the outgoing IAB members, each of whom received a plaque as an expression of the IAB’s gratitude for their service.
MPLS Turns 12
The IETF 74 technical plenary focused on MPLS. The session was intended to analyse MPLS, offering a case study involving the creation and operation of a successful protocol as well as the lessons learned. Four speakers enumerated MPLS’s benefits and its impact on the overall Internet architecture, including its effects on higher-layer protocol/application operation and delivery. The speakers represented operator, industry forum, and vendor perspectives: George Swallow of Cisco Systems, Tom Bechly of Verizon Business/MCI, and Kireeti Kompella of Juniper Networks. The discussion was introduced and moderated by Loa Andersson and Andrew Malis, who asked the panellists what they would have liked to change in the development of MPLS.
The speakers replied that overall, it is a good protocol, even though a few things could be improved. According to Tom, the diagnostic tools “trail the developments.” He said that having them sooner would have helped. Kireeti pointed out that a lot of machinery was put in the Label Distribution Protocol that is not used today and that complicates the implementation of the protocol. “There could also be more interoperability,” Kireeti said. “Interoperability tests take place in private networks, but there is not a lot of interservice provider interoperability.”
When asked about the fact that there are two protocols, Kireeti responded that having two protocols “is painful all around.”? This is true, he said, for the specifiers, the implementers, and the people who do interoperability tests. It was not a choice we made lightly. Learning from this experience and applying it to the deployment of IPv6, for instance, Kireeti commented that “Seeing so many variants of trying to get IPv4 and IPv6 to interoperate is amazing. I believe we need one solution and should deploy that.”
George added that the success of MPLS is to acknowledge that this is an IP world.
“Now we need to recognize that there needs to be a transition to IPv6,” he said. “In MPLS we are lucky though: we don’t have to deal with hosts.”
IAB Open Mic
The open-mic portion of the technical plenary included a discussion about the architecture of the Internet and how it has changed over time. Keith Moore expressed concern that there is no longer a set of shared assumptions. He said he’s hearing a lot of proposals that would violate the original set of principles of the Internet, and he’s wondering whether there’s a process that would bring us back to that state.
To illustrate that the Internet and the underlying assumptions have indeed changed, one participant pointed out that a number of young gamers, who were in San Francisco during the week of the IETF for a gaming convention, were surprised to hear that network address translation (NAT) was not an original assumption of the Net from the start.
Most of those who participated in the open-mic session as well as IAB members agreed with Keith, though most said they felt differently about the seriousness of the problem. Kurtis Lindqvist said he believes that the reason for the changes evolved because the Internet has been so successful. And that, he said, is because the Internet allows people to develop new applications and to make money. “It is an interesting observation, but I can’t decide if this is really a problem,” he said.
Scott Brim spoke to the fluidity of the Internet’s architecture, saying that it is even more fluid now. “There is a lot happening,” he said. “These are interesting times.” He agreed with Tony Hain, who earlier cautioned participants to “be careful that we can take out the stuff we put into the network once we solved the basic principles.”
Others felt more strongly that a shared set of assumptions is important and that one needs to make sure they stay consistent. “All of us struggle over how many of the original assumptions we can recover,” said Dave Oran.
The last part of the open-mic session addressed the work of the IAB and how it can be made more comprehensible and transparent. Most participants agreed that because the IESG is involved in operational issues, its work seems much clearer than that of the IAB. The role of the IAB is more difficult to grasp. Olaf encouraged people to speak up or to send suggestions to the IAB mailing list at email@example.com.
It is usually helpful when IAB members attend birds-of-a-feather or working group meetings. IAB statements or documents, too, are seen as useful contributions. “An IAB statement can have a lot of weight in the outside world,” said Alain Durand, even though, more often than not, the IAB is simply distilling ideas shared by the community, which is an important part of the work of the IAB.