|There's no way to make it more elegant!
|
|The SMPP specs dont allow for anything else. Nice would be to stop sending if we |got "throttle" and resume when we get some kind of "resume" message but there is |no such message. Kannel does the only valid thing which is wait a bit and try |again.
Adreas
 
Agree with you that there isn't in the protocol, however rather than the current 'sleep x and then try again' [bash head on wall, repeat] strategy that we use, we could use a more elegant method.
 
Something like the 'sleep for x, try again, if still errors, x = x + 1millisecond, repeat' until we get to the right automatic stage where we no longer get the errors, sleep automatically the right speed between messages and are therefore running at optimal speed for the connection.
 
This way with Rory's example, his application doesn't need to care, the box will just auto-throttle to the right speed for that connection. I've not tried this however if the SMPP PDUs come back as regularly as we need to slow down, then it should work
 
BTW, if you wrote the current code, I'm not having a go at it :-)
 
Alex

Reply via email to