←back to thread

BBR, the new kid on the TCP block

(blog.apnic.net)
360 points pjf | 2 comments | 09 May 17 08:56 UTC | HN request time: 0.458s | source
Show context
notacoward ◴[09 May 17 11:34 UTC] No.14299101[source]▶
It's almost irresponsible to write an article on this topic in 2017 without explicitly mentioning bufferbloat or network-scheduling algorithms like CoDel designed to address it. If you really want to understand this article, read up on those first.

https://en.wikipedia.org/wiki/CoDel

replies(5): >>14299422 #>>14299475 #>>14299613 #>>14301482 #>>14304236 #
1. apeace ◴[09 May 17 16:36 UTC] No.14301482[source]▶
It does mention buffer bloat:

> Larger buffers can lead Reno into a “buffer bloat” state, where the buffer never quite drains. The residual queue occupancy time of this constant lower bound of the queue size is a delay overhead imposed on the entire session.

replies(1): >>14305516 #
2. notacoward ◴[10 May 17 02:29 UTC] No.14305516[source]▶
Yes, it mentions bufferbloat once, but IMO the subject requires far more than that. I would suggest at least introducing the concept of active queue management, maybe set some of the developments in that closely related field - e.g. RED, SFB, CoDel - on the same timeline with Reno, Vegas, and BIC. Otherwise it's like trying to teach someone driving without mentioning stop signs and traffic lights.