Skip to main content

Filter by topic and date

Filter by topic and date

Reflections from IETF 97

3 Jan 2017

I wanted to send a post summarising my thoughts of the discussions at IETF-97. We had 1042 people from 52 countries on site in Seoul, very active development on a number of fronts, and I thought a successful meeting! (Apologies for sending this post out late, other tasks, holidays, etc intervened.)

The meeting was supported by our host Huawei and co-hosts CNNIC and KISA (Korean Internet & Security Agency), and a long list of sponsors. Thank you for your support!

The topic of the meeting was of course Internet tech and its evolution. The two most active discussion topics were the increasingly serious Denial-of-Service attacks that we are seeing, and the development of a new transport protocol, QUIC, as an alternative to TCP and TLS, especially being more optimized for HTTP/2 usage.

The most recent Denial-of-Service attacks involved a number of compromised Internet of Things devices attacking DNS infrastructure. The IAB had organised a discussion of these attacks as an example of a more general concern: the addition of millions of new hosts has the capability to overwhelm the Internet infrastructure when those hosts misbehave. There are ways to mitigate the attacks, but not without impacts in other ways — such as finding it necessary to deploy your services on large providers.

At the very least, I think it would be beneficial for the IETF community to continue to call attention to folks that the minimum bar when introducing a large number of devices (or any device) to the Internet includes things like automatic software updates and avoiding default passwords. I used to think this was so obvious and it needn’t be said, but I’m not so sure anymore. Nevertheless, the area for us to have an impact is improving on defence and mitigation mechanisms.

You can watch the video from the IAB plenary discussion here:

The IETF recently chartered a working group to specify QUIC (Quick UDP Internet Connections). This new protocol combines the TCP and TLS layers, is typically implemented in user space rather than kernel space, and aims for faster connection setups using resumption, integrated security, and capabilities to evolve the protocol faster (not being in the kernel).

A previous version of the protocol is also already in relatively wide use at Google, and was taken as a starting point for discussion in the working group.

I’m quite excited about this development, and eager to see where it takes us, and it seems that I’m not alone:

Once again the IETF Hackathon was running the weekend before the IETF. I thought it was outstanding to see large student groups among the participants. A student team from SungKyonKwan University worked on the Interface to Network Security Functions (I2NSF) framework, for instance. With jackets made for the event!

There was also a second large student team – on the other side of the world! The team from Université Catholique de Louvain worked on Multipath TCP, but much of their team did their work from back home in Belgium. They won the best of show award:

The Hackathon demos can be viewed in YouTube:

All videos from the sessions, interviews, etc from the meeting are available as a YouTube playlist. The official proceedings with slides, minutes and everything else are on the IETF website. See also the blog post on routing area outcomes from IETF 97, and the blog post from Srimal Andrahennadi from his experiences in participating as an ISOC fellow at the IETF.

So, what’s next? The work has already been going on in the mailing lists and virtual meetings after IETF-97, and will intensify as people come back from holidays. IETF’s official work happens on the mailing list. Our next meeting is at the end of March in Chicago, hosted by Ericsson:

Share this page