Hello!

Recent experiences testing sipxbridge has convinced me that we need a
SIP regression test tool.  We seem to go through a lot of testing
against actual equipment ( phones , gateways, ITSPs). This is
unavoidable. However, as development progresses it becomes pretty hard
to repeat all the scenarios as it is all manually done.  What makes
the problem hard is that each piece of equipment that we test with has
its own peculiarity. As I move forward as a developer, I want to be
able to test against all the previously successful tests but that is
pretty hard to do without configuring the actual equipment. Ideally, I
would like a tool that can capture the interactions at the endpoints
of interest and be able to replay these interactions at a later time.

The purpose of this mail is to discuss some ideas for such a tool. To
that end, I would like to kick around some ideas I am thinking about.
Here is an operational scenario:

A component developer or tester :
- Runs a scenario consisting of actual phones/gateways.
- Runs tcpdump while the scenario is running.
- Records the user names that were used in his test scenario and
endpoint (phone/gateway/ITSP) IP addresses that were used for running
the scenario.

Another developer who would like to run the test proceeds as follows:

- The captured trace provided by the component developer is provided
to the sip test tool (that we are developing) along with the IP
addresses of the endpoints that are the source of SIP signaling ( i.e.
the external stimulus ). This has been previously recorded by the
component developer.
- The test tool defines and configures pseudo users on the SUT for the
purpose of testing.
  These pseudo users are mapped to the test user identities defined by
the component developer above.
- The test tool analyzes the tcpdump and figures out the sip (client)
Transactions and Dialogs that were originated at the IP addresses
specified above.

Verification proceeds as follows.  For every client transaction
generated by the tool, the test tool does a causal analysis of the
signaling activity at the other test endpoints to determine what
requests must appear at those endpoints :
- The test tool runs the scenario by starting client transactions at
the appropriate times as indicated by an analysis of the pcap file.
- The test tool logs the messages seen by the other simulated
endpoints and analyzes whether these are received in the same causal
order as the previously analyzed tcpdump.

Does this all sound feasible?

As you (developers and testers) are going to be the primary audience
for this tool, I would like to get some early feedback on these ideas.

Than you for your thoughts in advance.


Regards,

Ranga.


-- 
M. Ranganathan
_______________________________________________
sipx-dev mailing list [email protected]
List Archive: http://list.sipfoundry.org/archive/sipx-dev
Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev
sipXecs IP PBX -- http://www.sipfoundry.org/

Reply via email to