From Lab to RFC: A PhD Student's Journey through the IETF

IETF Blog: While the IETF community is often thought of as nearly entirely composed of individuals from industry, more than a fifth of IETF participants work in academia, according to the latest IETF Community survey. What was your initial motivation to participate in the IETF?

Martine: During my undergraduate years, I wrote the network stack for RIOT, an IoT operating system, which eventually became my Master’s thesis project. For this, I had to read many IETF documents and participate in plug tests—also known as interoperability tests. Back then, dedicated plug test sessions happened in meetings right before an IETF week started. These were similar, but smaller, to what the IETF Hackathon is now. 

This is how I first got exposed to IETF and got hooked. From then I tried to participate at least once a year at one of the IETF meetings in addition to participating in mailing list discussions. Luckily, the research chair I work with was very supportive and I was able to find a PhD topic that relates to IETF work.

IETF Blog: How have you balanced your work in the IETF with your academic research, and to what degree have they been connected with each other?

Martine: My IETF work provided a good foundation and motivation for my research. I started designing what became RFC 9953, DNS over CoAP (DoC), because we needed a solution for encrypted DNS in a LoRaWAN (see also RFC 8376 and RFC 9011) deployment, the scenario for our research project back then. In previous research, we found that simply fragmenting messages in such scenarios would lead to significant performance penalties, so existing DNS over various protocols (HTTPS, QUIC, or (D)TLS)—which we call “DNS over X”—proposals were not an option. 

Starting from our initial proposal, we then empirically compared the performance of DoC with other DNS over X proposals. The results enabled us to improve our Internet-Draft even more. Since we do not only do networking but also systems research, we build the systems that we analyze. Therefore,  providing “running code” feels quite natural to me. Based on our improvements, I was able to identify further open problems, more drafts spun off, which provided more opportunities for evaluation, and so on.

IETF Blog: Has there been anything unexpected in your experience in the IETF over the years that has related to your research?

Martine: Some people claim academia moves slow, but I was surprised to learn that IETF can even move slower. Research articles take months to complete, RFCs years! 

Jokes aside, both proper research and engineering take time. And to be honest, sometimes it also takes years to publish a research article. What I did not expect, though, is that the IETF community is sometimes even more obsessed with finding the right solution and presenting it clearly. And getting detailed, constructive feedback from people who are not part of your direct peer group is also more common in the IETF. 

IETF Blog: What would you say to other students or academics considering how or whether to get involved in the IETF?

Martine: If you have the opportunity: Do it! It is great to meet and get feedback from people who define the fundamentals of Internet services and applications. If your research can contribute to this, it makes your PhD journey more satisfying despite the “little” extra work. IETFers are eager to work on Internet-Drafts together. And from the IETF work, new research opportunities may arise.

Participating in mailing list discussions is a good start but you should also try to attend the meetings onsite, at least once a year. There are ways to get support based on the Internet Research Task Force (IRTF) Travel Grant program (thanks!!), but also some research funding agencies understand that Internet standardization matters (thank you German Ministry of Education and Research). 

IETF Blog: Now you are nearly finished with your PhD–congratulations! Do you have a view about how you might continue to participate in the IETF?

Martine: Sadly, nothing concrete yet. For now, I want to do something outside of academia, but the current economic and political situation seems to make it difficult to find something that fits my profile and would also be safe for me. If your readers need someone in their team who wants to push constrained IoT protocols, such as 6LoWPAN, SCHC, or CoAP, forward with over 10 years of experience in working in both Internet standardization as well as Open Source development, they are welcome to get in touch with me.