Dave Täht, influential network engineer, has passed away

[-]

J_tt@reddit

I briefly hosted a VM for Dave and emailed back and forth with him, sad to hear the news, he was an amazing asset to the community and will be missed.

[-]

Portbragger2@reddit

that's sad and somewhat depressingly ironic considering i just successfully set up fq_codel on my pfsense box last week.

r.i.p.

[-]

Damn. Dave once appeared out of nowhere to answer a question I had on the Arch forums. He was always active on reddit and the Openwrt forums too — a tremendous resource. Those kind of direct interactions are one of my favorite things about FOSS.

I've always found the bufferbloat community's advocacy to be admirable. We all know the labor of FOSS maintainers is underappreciated, but many of course continue because they enjoy hacking. Dave Täht also did the tedious work, like engaging with the FCC in the bufferbloat NOI response in 2023 https://www.fcc.gov/ecfs/document/12010651418616/1. I figure that's not something you write because it gives you intrinsic joy, but because you think the world could be a better place and you're working to move it there.

[-]

IBNash@reddit

RIP mate, thank you for your gifts to Linux 🙏🏽

[-]

FerryCliment@reddit

Feeling that in the next few years we will get some more OGs leaving us, and its quite sad, I'm near my 40's and I was not that much in IT when I was young, but I've read lot of contributions Täht and other have made, something that today we take for granted, but... have a name behind these ideas, configs or products.

[-]

EmilyActually@reddit (OP)

This is really sad news to wake up to. I was only reading some old posts from Täht the other day, on the subject of active queue management (AQM) algorithms, over on the OpenWrt forums (he made a lot of good contributions to discussions over there, among other places on the Internet).

For those unaware, Täht was a key figure in spreading awareness of the issue of excessive queueing (bufferbloat) in modern Ethernet networks, especially across Internet connections of wildly-varying links speeds (whether xDSL, DOCSIS, fixed wireless, or cellular), along with solutions to tackle this issue.

Täht also fought against bufferbloat to make Wi-Fi more fair and responsive for all the wireless devices out there. Both him and Jim Gettys, along with other key people like Kathleen Nichols and Van Jacobson, were instrumental in the development of an AQM algorithm that would become CoDel - short for controlled delay, and pronounced as "coddle" - and later down the line, developments like fq_codel and the CAKE algorithm for more robust traffic control.

I still remember those pandemic times back in the early 2020s, back when so many of us were stuck at home, and so many of us were reliant on stable Internet connectivity to work from home. And it's largely thanks to algorithms like fq_codel that kept my network super stable, even if my whole family was home, all using the internet at the same time as me.

If this is your first time hearing about bufferbloat, I would recommend reading through the Bufferbloat wiki. There is a lot of good information there to cover the bases, along with ways to mitigate the issue.

Looking back, we owe so much to Dave Täht, for all his contributions. He will never be forgotten. I hope he rests in peace, and I wish all his family and close friends all the best, following this tragic news.

[-]

wunderspud7575@reddit

I never knew Dave, and never interacted with him directly. But I read his writings, and followed his work closely, and he was truly a unique and brilliant mind.

One of my approaches to self learning is to identify great engineers and follow their work and learn from their design choices and approaches. I am sad that there won't be more from Dave, but what he leaves behind is a masterclass in engineering.

RIP.

[-]