On 02/04/16 12:19, Marko Cupać wrote:
If you need something more advanced, like granting minimum bandwidth
with the ability to borrow from other queues up to maximum bandwidth
when other queues are empty, I am afraid you won't be able to do it
with current queueing system. However, I'm still hoping one of
these days someone will prove me wrong.
I'm a little curious about what you're basing this assertion on?
As in, can you point us to benchmarks or other tests that indicate that
hfsc queueing (which is the algorithm underlying the post-5.5 traffic
shaping) with variable allocations does not work?
Child queues with variable allocations (min and bandwith not equal to
max or burst) can never exceed the parent queue's allocation, but using
the min, max and burst parameters as documented in the queueing section
of pf.conf(5) should indeed influence observed results /when one or more
queues approach a threshold that triggers shaping/.
If you can come up with measurements that show the mechanism is
defective, I'm sure a bug report would be appreciated.
--
Peter N. M. Hansteen, member of the first RFC 1149 implementation team
http://bsdly.blogspot.com/ http://www.bsdly.net/ http://www.nuug.no/
"Remember to set the evil bit on all malicious network traffic"
delilah spamd[29949]: 85.152.224.147: disconnected after 42673 seconds.
