Dave Taht

Achieving a consistently low latency, less annoying internet

dtaht has 0 patrons.


Hi. I'm Dave Taht. You probably don't know me, but it is highly likely every day your packets now flow through algorithms my project (bufferbloat.net) helped design, reducing network latency by orders of magnitude, and making your internet experience better. Some are pretty well known now, fq_codel (RFC8290), "smart queue management", and most recently "cake", although I've had a hand in nearly every new tech that reduces latency over the internet, like packet pacing. There are billions (not kidding) of deployed implementations of this stuff now, silently, doing their job, for you, but with an impact so diffuse, that you never knew they were there.

If I had a penny for every packet dropped or rescheduled by one of "my" algorithms, I'd be a billionaire in a hour (not kidding, either)! But, instead, I've been at this over a decade now, mostly living below the poverty line, trying to make sure these algorithms went into every router, server, and client on the internet, patent free, and open source. We seem to be succeeding! But theory work... that takes a decade to deploy - doesn't pay well. I exhausted my own savings early in the project, and I've turned down gobs of money from major corporations that would have immediately found a way to patent and profit from my work, because I felt the world needed a better internet, and I felt I owed Vint Cerf, Jim Gettys, ESR, Dave Reed, and so many others, something in return for their enormous contributions to the internet and my career.

Most of my work in the last few years has been more political, than technical, although I've kept a hand in on improving wifi, and my most active project is libreqos.io (which attempts to get the latest latency-busting algorithms into ISP's hands so they can deploy easier). In the coming year, I'd like to try to refocus more on explaining the why, at a level of detail that can be carried forth by the next generation of researchers and internet experts. I'm terrible at certain things (like graphics), and would like to find or hire a volunteer that could turn text and math into pictures, and animations, that explain that "why" better. (I'm also terrible at getting grant money although I've had various sponsors in the past like nlnet and comcast research)

I'm getting old, my mind is getting crunchy, and I'd really like a chance to write more stuff down, before retiring, oh, say, 3 or so years from now. I'm not in a position, financially, to do that, but I'm going to have to quit for health reasons by around then, the way things have been going lately (covid took a terrible toll on me). I also have a couple new things, that I think would be useful to develop - and give away - , that would also take more than a few years to deploy, that I'd like to work on, or find folk willing to work on.

Linked Accounts

dtaht owns the following accounts on other platforms:


dtaht joined 3 months ago.

Income Per Week (in US Dollar)

Number of Patrons Per Week