Re: Patch: To add timeouts to SMPP connections

> >From SMPP 3.4 specs ...
>
> "The maximum number of outstanding (i.e unacknowledged) SMPP operations
> between an ESME and SMSC and vice versa is not specified in the SMPP
> Protocol Specification and will be governed by the SMPP implementation
> on the SMSC. ... as a guideline .. no more than 10 SMPP messages are
> outstanding at any time. " (2.5.1 for example)
>
> If you were going to hard-code a X - better make it 10 (use #define too
> by preference). Otherwise I agree with Alex .. the 'governed by SMSC'
> bit implies .. many and varied.
>
> Ultimately however review sections 2.9 and 7.2 of the specs regarding
> SMPP Timers as implementation guide.
>
> Just my NZD0.02

I had noticed this, but it also refers to any outstanding primitive, such as
sending messages. It's possible that the SMPP server could be hundreds
behind on confirming message delivery by the time it's ten behind on
'ping's. Though, for light operation, this would not be the case.

Without writing hooks to confirm each kind of outstanding message, this
figure will probably have to be guesstimated.

I'm beginning to see that there probably is a good argument for having this
as a variable, and would definately like to take up Alex Judd's offer to
rework the patch to use the Kannel configuration mechanism.

Aaron