[asterisk-biz] asterisk-biz Digest, Vol 104, Issue 9

[Rifat Rahman]

Hello Gerrit,

> I personally would not implement this in kernel space.

Me too. But sometimes we have to follow orders. As I was new, and had
to follow orders, I did.
And the project ending as under my lead (!) as it ran so far and I was
able to perform critical
R&Ds about kernel space packet mangling..

> I would have chosen a softswitch platform where this can be easily implemented as a module.

This is the choice. Kamailio + rtpproxy is well as proxy I think. This
is in use here. But the target was to
decreasing a hop, to speed up the encryption, padding and ptime
modification process, and improvement
of quality.

Whatever, I learnt a lot about linux kernel and it changed my mind in a
ridiculous way. At least I got the chance of kernel programming. As it
is showing the horizon,
so I have the pressure to make it work. But from the beginning, my
suggestion was to observe carefully
the softwares, tuning them properly, and if possible try to improve
them rather than writing the whole
logic in kernel space. And it was really horrible at first.

Rifat