[Bloat] In loving memory of Dave Täht <3

Livingood, Jason jason_livingood at comcast.com
Wed Apr 2 09:59:57 EDT 2025


Very sad news indeed! I had the pleasure of working closely with Dave for 15 years. He was generous with his time and had a unique way of bringing people together to make the internet better for everyone!


I had to go down memory lane to recall when I first really started working with him. It may have been around 2010 or so. In 2012, I started sending funds his way via my day job to help him and his merry network of collaborators work to develop the CoDel AQM. 


Funding him was not necessarily easy, as Dave had a unique way of working and was best when he had complete autonomy and only loosely outlined goals - typically hard to sell in a big company. But he could make things happen, so it worked. And I knew when he started complaining about maintenance needs on his boat, or the need to recruit a new person to the project, or about a great new (and practical!) idea, that it was time to top up his funding. ;-)


That initial CoDel support in 2012 was extended to underwrite work on his idea to develop RRUL, the first real working latency test that I can remember (https://www.bufferbloat.net/projects/bloat/wiki/RRUL_Spec/ <https://www.bufferbloat.net/projects/bloat/wiki/RRUL_Spec/>). He was also helpful in introducing me to Simon Kelley, developer of dnsmasq, so we could underwrite some IPv6 features in dnsmasq (and Dave convinced Simon to come to an IETF meeting to help gather requirements and meet folks). 


Dave got CoDel working, so we developed a compelling demo of CoDel on a DOCSIS network (via a CeroWrt-based router connected to a cable modem) and brought him along to IETF-86 in March 2013 in Orlando - see interview with Dave at https://youtu.be/NuHYOu4aAqg?si=p0SJHLNpp_6n7XP9&t=195 <https://youtu.be/NuHYOu4aAqg?si=p0SJHLNpp_6n7XP9&t=195>. 


From 2014-2017, I was able to make additional financial support happen for him, so he could do R&D into how to improve buffer bloat in WiFi network links and equipment, a project he called "Make WiFi Fast". In 2020-2021 and 2024, I found funding for his work again, this time to work on accelerating AQM adoption in the real world & work related to the CAKE AQM.


Thanks in part to my longstanding collaboration with Dave, tens of millions of DOCSIS users in our network have AQM and thus far better network responsiveness. The same is true for AQMs he worked on, CeroWrt, LibreQoS, and other projects. He succeeded in his goal to make the internet better for everyone!


We will miss you, Dave!


Jason













More information about the Bloat mailing list