Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Improve performance on long fat networks (LFNs) #2910

Merged
merged 1 commit into from
Mar 22, 2024

Conversation

matt335672
Copy link
Member

Fixes #2905
Probably fixes #1625
Fixes #1483
Fixes #1026

On Linux, the TCP send buffer size is increased to 32768 if it is less that this (which it normally is). This however has the effect of disabling dynamic buffer sizing, leading to a maximum available bandwidth of

max_bandwidth = 262144 (bits) / round_trip_time (secs)

This is not noticeable on a LAN with an RTT of around 0.5ms, but very noticeable on a WAN with an RTT of 0.25s.

Comments in the config file and manpage in this area are improved, as is the logging if the parameters are actually set.

This change has no impact on default FreeBSD behaviour. On this platform the default TCP buffer size (net.inet.tcp.sendspace) is 32768.

On Linux, the TCP send buffer size is increased to 32768 if it is less
that this (which it normally is). This however has the effect of disabling
dynamic buffer sizing, leading to a maximum available bandwidth of

max_bandwidth = 262144 (bits) / round_trip_time (secs)

This is not noticeable on a LAN with an RTT of around 0.5ms, but
very noticeable on a WAN with an RTT of 0.25s.

Comments in the config file and manpage in this area are improved, as
is the logging if the parameters are actually set.
@matt335672 matt335672 merged commit fb34d74 into neutrinolabs:devel Mar 22, 2024
13 checks passed
@matt335672 matt335672 deleted the fix_lfn_performance branch March 22, 2024 12:08
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
1 participant