Monthly Archives: March 2013

Running Code at IETF-86

bloat

Our meeting in Orlando ended on Friday. I thought it was a very successful meeting, and brought up many new topics that we should pursue. I will talk about some of those topics in this blog in the coming weeks.

After the meeting ended, I talked to some of the people who were coming into to the IEEE meeting that is taking place in the same hotel right after IETF. While our two organisations are different, we share some of the same participants, and some IETFers stayed in Orlando for two weeks. Our organisations also share many of the same visions about how standards should be defined in an open manner, and face many of the same challenges in our work. I learned a lot from my discussions with IEEE.

But back to the IETF. I wanted to write about some of the technical work that was going on during the week, but then I realized that it might actually be better to invite people who were actually doing the work. With this in mind, I want to introduce Chris Griffiths from Comcast. He talks about some of the testing and demos that were going on in our new Bits-N-Bites program. As you know, in the IETF we like to focus on running code, and I thought Chris’ story highlights this nicely.

Jari Arkko, IETF Chair

—-

I was asked to discuss some of the work that the IETF is doing with their Bits-N-Bites program, and some of the demos and lab work we did at IETF86 in Orlando this past week.

As we were preparing for the Orlando meeting, we talked about ways to make the Bits-N-Bites program more dynamic and provide a place for the IETF community to perform experimentation and get running code in a production like network setting. We shipped one of our Cable Modem Termination Systems (CMTS), which run on the Comcast production network down to Orlando and turned up a dual stack environment for demos and running code testing.

One example of this was the Homenet Working group running code examples based on the Homenet Architecture draft and the HIPnet draft. It was very exciting to see the collaboration between the many groups involved with setting up these Homenet demos, and great to see running code in action.

homenet

Another demonstration we performed live at Bits-N-Bites was an example of how to deal with a problem called Bufferbloat which is a problem discovered by Jim Gettys and worked on by a number of Open Source teams across the Internet. It was a very dynamic demo that showed the impact of bloated buffers on web performance and real time communication applications. We were able to show both bloated and dynamic buffer examples to the Bits-N-Bites audience, and it was well received. It was a great example of solving Internet scale problems with running code.

As we close up IETF 86 in Orlando, I hope that the IETF continues with the lab experiment we started here this week. It was great to see first hand examples of running code and teamwork that really energized the IETF community, which I think we need more of.

Thank you and I look forward to seeing everyone in Berlin for IETF 87.

— Chris Griffiths

Welcome to IETF-86!

Orlando-Florida-Skyline

I would like to welcome you all to Orlando, where the 86th IETF meeting starts on Sunday! All the details of the meeting’s agenda and other important information is available from the IETF web page.

I arrived in sunny Orlando today. The staff and volunteers have already been on site for some time, working hard to prepare for the meeting. The meeting network is coming up, and everything looks to be ready for the meeting to begin.

aps

Comcast and NBCUniversal are hosting this IETF.  They have also arranged an exciting social event on Tuesday evening at an exclusive location inside Islands of Adventure® at Universal Orlando® Resort. Bright House Networks is sponsoring the network connectivity. Polycom is a meeting sponsor. I would like to thank our hosts and sponsors. Their big contribution makes it possible to have a good meeting here. Thanks!

I would also like to highlight a few items in our meeting agenda that were particularly interesting for me at least:

  • On Saturday, March 9th, volunteers meet in a Code Sprint event, working together to build tools to enable IETF work more efficiently. Please join us!
  • On Sunday, March 10th, Bob Hinden from the administrative oversight committee will explain IETF finances in Caribbean 6.
  • On Tuesday and Thursday, the RTCweb working group meets to discuss this new technology that is enabling browsers to act as phones.
  • Software defined networking topics are discussed in the I2RS working group and the SDNRG research group.
  • Many working groups discuss technology necessary to build the Internet of Things. For instance, on Tuesday the CORE meets to discuss its lightweight protocol to replace HTTP for small devices. Other working groups to follow include ROLL and LWIG.
  • Improved authentication mechanisms for HTTP are being discussed on Tuesday in the HTTPAUTH session.
  • The ECRIT working group defines mechanisms necessary to support emergency communications. Their meeting is on Wednesday.
  • On Monday evening, the whole IETF meets in the technical plenary to discuss the technical and regulatory impacts of decommissioning the plain old telephone system or POTS.
  • The LMAP BOF is discussing large-scale measurements to look broadband quality across large areas and providers.

Jari Arkko, Incoming IETF Chair

IETF Challenges

nat-small

The previous article talked about how exciting and important the work at the IETF is. And it is. But there are also challenges, both for the Internet as a whole and for us at the IETF.

The Internet keeps facing both technological as well as societal challenges. The fast growth of the Internet makes scalability very important. New applications push the limits in other ways. And the enormous importance of Internet communications in our personal lives and economic activities makes the Internet also a part of legal and political interests. Retaining an open, one Internet while tackling many of these challenges is of utmost importance.

I am sure we will discuss the above at length in the future, in the IETF and elsewhere. But I wanted to focus on this article a little bit more narrowly on the IETF.

Here are some issues that need attention:

  • Addressing the needs in important technical areas, such as real-time communication, the Internet of Things, or IPv6 deployment. Our highest priority is to produce timely, relevant, and high-quality standards. As long as the industry and users adopt our solutions, then we are on the right path. On many of these areas there is plenty of work left, however, as well as opportunities to take on more work.
  • Identifying the new technical challenges that face us, such as power constraints (be it in datacenters or small devices). What are these challenges?
  • Evolving participant base. As our topics change over time, so does the set of people with expertise on those topics. For instance, in the area of emergency communications we have to find ways to interact with people from regulatory agencies. Similarly, the IETF has become very international, with document authors from 60 countries. But there is still work left to make our organisation and leadership even more international and more diverse.
  • Dealing with the age of “permissionless innovation”. Internet technology enables building applications in an easy manner, by anyone. And usually without any effect on the underlying Internet protocols – the part that IETF is about. And even where there is an impact, there is often an interesting tussle about what aspects need to be standardised. E.g., fully specified real-time communication protocols vs. frameworks such as WebRTC that can be used to build solutions. Finding the right balance between these types of approaches is important.
  • It is not always easy to start new work at the IETF for various reasons. And “the end-to-end delay”, time that it takes from proposing a BOF to having a WG and getting an RFC out is still very long. Even if we have improved how we handle specific smaller tasks, like approving an RFC, building an entirely new specification for a new problem takes a lot of time.
  • The IETF process puts more weight in the final stages, and the role of the IESG is quite central. It would be better to push more of the review work to earlier stages. At the same time, this would reduce the load on the Area Directors. It is not always easy to find Area Directors willing to devote enough time to the task of being in the IESG.
  • IETF’s process documentation is in the need of revision, in some cases even to bring documentation up to the state of currently used procedures.

We will see how these issues can be tackled. I do have an idea about some of the principles that we should employ in that, however. The first is continuous, incremental improvement. The second is transparency, keeping everyone informed about what is going on and calling for feedback. The third one is to focus. Fourth, running code and rough consensus. Code, interops, engineers. Publish and prune RFCs easily.

But enough about my thoughts. What I really want to know is what do you think. What is troubling you at the IETF or Internet technology? What new technical challenges do you believe IETF should tackle? If you have comments, send them directly to me or post to the IETF discussion list.