Brian Carpenter, a distinguished engineer with IBM, has taken over as chair of the IETF, the Internet’s foremost standards-setting body. Carpenter is the sixth person to chair the IETF since its founding in 1986. Network World Senior Editor Carolyn Duffy Marsan had an exclusive interview with Carpenter at the group’s recent meeting in Minneapolis. Here are excerpts from their conversation.
Brian Carpenter, a distinguished engineer with IBM, has taken over as chair of the IETF , the Internet’s foremost standards-setting body. Carpenter is the sixth person to chair the IETF since its founding in 1986. Network World Senior Editor Carolyn Duffy Marsan had an exclusive interview with Carpenter at the group’s recent meeting in Minneapolis. Here are excerpts from their conversation:
What is your role at IBM?
I handle internal coordination regarding the IETF. I’m also working on product development strategy for networking. The IETF chair is a full-time job, but I am keeping a few customer projects. I don’t want to be too removed from reality.
What do you hope to accomplish as IETF chair?
My first goal is an internal goal for the IETF, but I would like to get the Internet Engineering Steering Group into a position of steering and not just processing documents. I would like the members of the steering group to get into more strategic issues of where we should be going in the future. Second, I want to improve our liaison relationships with other standards bodies, including the International Telecommunication Union, the Third Generation Partnership Project and the Open Mobile Alliance. I want these relationships to work in both ways.
What is the most important technical work going on in the IETF right now and why?
I still believe that getting IPv6 deployed is very important. I think the people who say we can go on with IPv4 are strategically wrong. Second, I think we have work to do in the security area. There is still more work to be done to have an overall security model. The other important work is [Session Initiation Protocol] and everything that goes with SIP. This is where the liaison issue with the ITU comes up because we’ll be using SIP to run regular telecommunications services over the Internet.
What will you do as IETF chair to promote IPv6?
The IETF’s main job on IPv6 is done, and it is now shipping in major operating systems. It’s now a deployment issue, not a development issue. The IETF needs to keep saying that IPv6 is a strategic technology and that it’s ready. There is already strong commitment by the U.S. Department of Defense, China and the 3G [Partnership Project] to IPv6. A small or medium-sized business probably has no urgent need for IPv6. But for the enterprises, sooner or later they will notice that all their operating systems and routers can do IPv6. Then they’ll notice that it’s not a big capital expense to move to IPv6. But it may be a big deal operationally.
Many IETF working groups are behind schedule. What will you do to encourage them to wrap up their work?
There’s no single answer. When you dig into why something is moving slowly, it might be pure sociology as in the working group dynamics or it could be a patent issue or a general technical difficulty or there could be something wrong with our processes. Other standards bodies meet their deadlines but they put out documents with certain items still open for study. These documents aren’t usable by implementers. We don’t do that. It can be frustrating for other organizations, but it has always been the IETF practice to wait until we have a standard that can be implemented.
What concerns you most about the health of the Internet infrastructure?
Where [Multi-protocol Label Switching] and [Generalized MPLS] go are clearly a matter of concern. We still have security issues at the infrastructure level, including securing [Border Gateway Protocol ]. Internationalization is also a very interesting issue because it’s a matter of great interest to the users. But it’s a matter of very little interest to some of the technical people because they view it as a presentation-layer issue.
There is a perception that more of the cutting-edge standards work is going to other groups, such as the Organization for the Advancement of Structured Information Standards and the Liberty Project. Is the IETF losing the cool work to rival standards bodies?
It’s clear that one of the things we need to do is understand the scope of what we do in the applications area. For example, in the area of Web services , a lot of standards work is done by the World Wide Web Consortium and OASIS. I don’t think that’s a point of tension. Everyone agrees that the IETF is the maintenance body for [Hypertext Transfer Protocol], while everything that runs on HTTP is at W3C. There’s some virtue in specialization for the standards bodies. So I don’t believe we’re losing the cool work.
Is the IETF still the Internet’s premier standards-setting body?
Yes. I believe it is for infrastructure standards. For Layers 3 and 4, we are the leaders. I don’t believe that’s in doubt.
Attendance at this meeting is down, from a high of 3,000 a few years ago to a little more than 1,000 now. How big of a problem is that?
What I’m hearing is that things are working much better than when we had 3,000 attendees and people were spilling out into the hallways. Now the people who come to our meetings are serious about the work we do.
Your predecessor, Harald Alvestrand of Cisco, spent much of the past two years involved in administrative restructuring. Do you think you will, too?
What I’m hoping is that we can fairly expeditiously finish administrative restructuring and get the new administrative oversight committee in place so it can become a self-contained entity. I will still be concerned about IETF workflow, IETF processes and IETF sociology. Part of my job is to make sure all of that runs smoothly.
The IETF chair is a huge job with no pay and lots of hassles. Why take on this job?
Partly because people were twisting my arm, but mostly because I think it’s important to grow the Internet for the benefit of the whole world.
AT&T, MCI and Sprint have traditionally been active participants in the IETF but they are all undergoing mergers. The regional carriers – SBC, Verizon and Qwest – have not been as active in the IETF. Will ISP consolidation hurt the IETF?
That’s a hard question. Larger companies tend to move more slowly to new technology, but when they move they really throw their weight around. The question is when these carriers will need to upgrade. The smaller carriers have tended to be end users rather than innovators in the technology. Once the acquisitions shake down, I can’t see how they won’t be interested in new technologies.
