Equivalent jitter-buffer for kernel space

[Myng]

Hello, 
since jitter-buffer is support for user-space only, eg: --jitter-buffer=200

I read some params like dtx, but unsure what is good way,  can anyone show me how to archive this but for kernel space? 

Thanks in advance


 

[Richard Fuchs]

There's no such thing at the moment. Kernel forwarding is a more or less direct in-to-out operation with only minimal processing.

Cheers

  --
You received this message because you are subscribed to the Google Groups "Sipwise rtpengine" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rtpengine+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/rtpengine/2f289f3d-1a8c-4927-8664-e441c62e81ddn%40googlegroups.com.

[Myng]

Thanks @rfuchs for your answer, 
I know that that kernel space offer better performance compare to user-space, but do you have any idea how much it better and especially in case transcoding? 

Regards, 
 

[Richard Fuchs]

On 23/06/2025 09.55, Myng wrote:
> I know that that kernel space offer better performance compare to
> user-space, but do you have any idea how much it better and especially
> in case transcoding?

There is no support for transcoding in kernel space. The difference in
performance would likely be negligible, because the load caused by the
transcoding process itself likely outweighs the overhead of having to
move data between kernel and user space (except possibly for certain
trivial transcoding scenarios, say PCMA <> PCMU).

For passthrough it's hard to give exact numbers as it heavily depends on
the hardware used, but a good rule of thumb is about 500 concurrent
calls per CPU core with user-space forwarding, while with kernel-mode
forwarding the CPU load essentially disappears (typically for example
from ~500% down to ~30% or so) and you're mostly limited by how much the
network hardware and the driver is able to push.

Cheers