Hi Lilin, 

You can set up your deployment to generate large amounts of SIP stress by 
adding some Clearwater SIP Stress nodes (how to do this is described in detail 
at 
https://github.com/Metaswitch/clearwater-docs/wiki/Clearwater-stress-testing). 

These nodes use a standard SIPp script 
(https://github.com/Metaswitch/sprout/blob/icscf/tests/load/call_load2.xml) 
which simulates a pair of subscribers registering every 5 minutes and making a 
call to each other every 30 minutes. The script runs on Clearwater stress nodes 
against the bono cluster in a deployment, each node accounts for 100,000 
subscribers, and multiple stress nodes can be set up.

Hope this helps,

Ellie

-----Original Message-----
From: Lilin [mailto:[email protected]] 
Sent: 30 May 2014 20:56
To: Eleanor Merry; [email protected]
Subject: Re: [Clearwater] Clearwater Test Scripts

Thanks Ellie! The information is very helpful ;)

I have identified at least, in stack.cpp as well as in stateful_proxy.cpp, both 
files have the data structure, pjsip_rx_data* rdata, which contains the SIP 
messages sprout receives.

Next I'm about to decide where to implement the feature of recording the 
start/completion timestamp of a call session and the RTP endpoints of both 
sides.

In the meantime, I also want to know if there exists a script that can do 
stress testing, i.e., establishing a few hundred simultaneous SIP requests to 
the bono proxy.

I understand that the existing test suite, clearwater-live-test, provides an 
option of repeating the call operation for a specified number of times. But 
these SIP requests are sent sequentially. Please advise a way of stress 
testing. Thanks!

Regards,
Lilin

On 29/05/2014 8:40 AM, Eleanor Merry wrote:
> Hi Lilin,
>
> You can find the details for building/installing each clearwater component in 
> the docs section of the repos - e.g. for sprout/bono this is 
> https://github.com/Metaswitch/sprout/blob/dev/docs/Development.md.
>
> The stable branch in each repo marks the latest release (currently Final 
> Fantasy), I recommend this branch unless you need the very latest code (which 
> is in the master branch).
>
> The 'stateful_proxy' module registers for receiving requests and 
> responses; it is responsible for handling requests and setting up the 
> necessary transactions. A description of the different modules, their 
> functions and their structure is at: 
> https://github.com/Metaswitch/sprout/blob/dev/docs/Bono%20Sprout%20Arc
> hitecture.md
>
> Hope this helps!
>
> Ellie
>
> -----Original Message-----
> From: [email protected] 
> [mailto:[email protected]] On Behalf Of 
> Lilin Zhang
> Sent: 27 May 2014 03:15
> To: [email protected]
> Subject: Re: [Clearwater] Clearwater Test Scripts
>
> Thanks Ellie,
>
> I plan to modify the source code to implement this feature (automated 
> script that records the start and completion time-stamp of a call 
> session, as well as the RTP endpoints of both sides.)
>
> If you could provide the information on the following, I'd be very
> appreciated:
>
> (1) Would you please provide me the instructions of installing clearwater ims 
> from source code, and/or the instructions of re-compiling clearwater ims?
>
> (2) I need to know which branch is stable that I shall use (master, 
> dev, stable, rtr_cache, etc?)
>
> (3) Which cpp file/functioin in sprout that handles the incoming call session 
> ( it would save me some time, if you could point it out ) Thanks!
>
> Regards,
> Lilin
>
> On 22/05/2014 7:21 AM, Eleanor Merry wrote:
>> Hi Lilin,
>>
>> We don't have a tool as you describe.
>>
>> Have you looked at our call latency stats though - this will give you 
>> average call latency over a 5 second period. Details of our stats are at 
>> https://github.com/Metaswitch/clearwater-docs/wiki/Clearwater-SNMP-Statistics,
>>  and there are some scripts at 
>> https://github.com/Metaswitch/cpp-common/blob/master/scripts/stats.
>>
>> If you need more fine grained results, then our logging code is in 
>> https://github.com/Metaswitch/cpp-common/blob/master/src/logger.cpp. This 
>> calculates the current timestamp every time it makes a log. If you do add 
>> this, please do consider contributing this back to Clearwater!
>>
>> Ellie
>>
>> -----Original Message-----
>> From: [email protected]
>> [mailto:[email protected]] On Behalf Of 
>> Lilin
>> Sent: 20 May 2014 22:05
>> To: [email protected]
>> Subject: [Clearwater] Clearwater Test Scripts
>>
>> Hello,
>>
>> I see that clearwater-live-test suite provides only binary outputs, 
>> such like PASS or FAIL for each scenario (Basic Registration, Basic 
>> Call, and
>> etc.) But I want to see a way so that the performance can be measured 
>> numerically, and done for each two-party call session (i.e., the 
>> measurement needs to be taken at the IMS server side, not merely at 
>> the testnode side.)
>>
>> I'm writing to see if there is already an automated testing script that 
>> records the start and completion time-stamp of a call session, as well as 
>> the RTP endpoints of both sides.
>>
>> I see that these time-related information (time-stamp for the sip messages - 
>> INVITE, 200 OK, BYE) and RTP endpoints already exist in sprout's log file. 
>> Hence there must be a function does this log-writing job. Which function is 
>> it?
>>
>> I'm thinking about to modify this function such that each time when 
>> it is about to write a log line (for the particular SIP messages - 
>> INVITE,
>> 200 OK, BYE), the function will also invoke a separate script (my
>> implementation) and record the time-stamps and RTP endpoints to another 
>> numerical measurement result file. Please let me know you opinion about this 
>> modification. Thanks!
>>
>> Regards,
>> Lilin
>> _______________________________________________
>> Clearwater mailing list
>> [email protected]
>> http://lists.projectclearwater.org/listinfo/clearwater
> _______________________________________________
> Clearwater mailing list
> [email protected]
> http://lists.projectclearwater.org/listinfo/clearwater

_______________________________________________
Clearwater mailing list
[email protected]
http://lists.projectclearwater.org/listinfo/clearwater

Reply via email to