16 ms·
This thing is cool. I saw a live demo at IETF 118 in Prague last month. It totally eliminates buffer bloat, which makes it awesome for video chat. I saw the
by toomim 3y ago
This thing is cool. I saw a live demo at IETF 118 in Prague last month. It totally eliminates buffer bloat, which makes it awesome for video chat. I saw the demo and was like "woah... I didn't think this would ever be possible."
It requires an additional bit to be inserted into IP packets, to carry information about when buffers are full (I think?), but it actually works. It feels like living in the future!
- ajb 3y agoThat bit is already there. L4S changes the meaning of the bit to allow a more accurate signal.
- toomim 3y agoYes, thanks for the clarification. IIRC was explained to me as "we put the last unused bit in IP packets to use, and get this great feature from it."
- ajb 3y agoYeah. It being the last bit (really the last codepoint in a 2 bit field) there was a big argument over it: https://datatracker.ietf.org/meeting/interim-2020-tsvwg-01/session/tsvwg https://datatracker.ietf.org/meeting/interim-2020-tsvwg-01/s... https://mailarchive.ietf.org/arch/msg/tsvwg/rXWRHAyGOuu_qOGM3JkpK9whdS4/ https://mailarchive.ietf.org/arch/msg/tsvwg/rXWRHAyGOuu_qOGM...
- fragmede 3y agoMore particularly, L4S is an advancement to the existing ECN (Explicit Congestion Notification) extension to TCP/IP, allowing for more advanced algorithms to cut down latency further.
- xorcist 3y agoThe main problem with ECN was the remarkably widespread behaviour by middleboxes that either cleared that bit or straight up dropped the packets. Maybe that situation has improved now?
- jlivingood 3y agoIt has definitely changed. Looking at measurements on this it seems there is only 1 major transit network doing that, and they are working to fix it.
- averageRoyalty 3y agoIt took a bit of searching, but I assume this is it? https://youtube.com/watch?t=4900&v=RWjbrXxpzVU https://youtube.com/watch?t=4900&v=RWjbrXxpzVU (1hr21m for anyone whom the time link doesn't work for) EDIT: Never mind, that's their hackathon recap. Still searching, this is not an easy conf to find talks for!
- toomim 3y agoWhat I saw wasn't in a talk; it was a 1-1 in the hackathon demos room.
- namibj 3y agoActually somewhat better, even: you can let it control the rate factor of your video encoder directly, getting perceptual fairness instead of simple naive bandwidth fairness.