[SIPForum-discussion] SIP testing

[Mark R Lindsey]

On May 11, 2011, at 5:12, Guido Negro <guido.negro at gradwell.com> wrote:

> Hi all,
> 
> 
> I am writing an app for system testing of a sip-endpoint. I'd like an opinion about some definitions I need to fix.
> 
> In my case a SIP test is a program that acts like a the phone and verifies that another SIP end-point behaves in a specific expected way.
> 
> Therefore I defined a SIP test failure : "a situation where a test is able to detect that the SIP end-point does not behave in the expected way".
> 
> That said, in my opinion, there are only 2 situations that could cause a so-defined SIP test failure:
> 
>     * a timeout expires while a test is expecting a specific request/response
> 
>     * a different response is received than the one expected (skipping 1xx of course)
> 



Or, perhaps the endpoint doesn't do the right thing. E.g., a phone returns 180, but makes no ringing noise. 

Or, perhaps the endpoint sends a request you didn't expect or want, like  a BYE before you're done with the session.

It's easy to be SIP protocol compliant with your given set of chosen RFC's and drafts en-vogue, but still get nothing useful done.