Laminar TCP for FreeBSD
hiren at strugglingcoder.info
Thu Sep 24 17:33:05 UTC 2015
While trying to debug TCP's behavior on certain loss situations at $work
we realized that even if we have this awesome modular congestion control
mechanism mod_cc(4), boundary between what stack should do and what cc
should do is not clear. For example, we currently have multiple
ways at multiple places where we adjust/modify congestion window based
on where we are in transfer and what RFCs we support. Downside it, it is
really hard to debug/analyze TCP performance in certain situations. Not
to mention the added complexity when trying to add a new feature or
modify existing one related to congestion control.
An interesting piece of work came out from Google that tries to address
this and more:
"The new framework separate transmission scheduling, which determines
precisely when data is sent, from Congestion Control, which
determines the total amount of data sent in any given RTT."
It'd be nice if we/FreeBSD can start looking into this as a long-term
TCP stack improvement project.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 603 bytes
Desc: not available
More information about the freebsd-transport