|
Home > Archive > BizTalk Server General > October 2004 > Biztalk Testing
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
|
|
|
| Hi All,
I am relatively new to BizTalk. We are working on a
project where the first step is to figure out whether
Biztalk caters to our needs.
I have gone through the performance white paper. But I am
not quite sure about how they measured the through put. We
want to run our own tests and see BizTalk's performance.
Please advise me how to calculate through put. We might
try to simulate the tests done in Performance Paper.
Thank you,
Rick.
| |
| Alan Smith 2004-10-22, 7:47 am |
| Hi Rick,
I have done a bit of basic performance testing on my current project.
The easyest way to test throughput is to use the file adapter with an
orchestration, and drop a bunch of say 1,000 xml files in the recive folder.
You can then use the XLANG/s Orchestrations and BizTalk messaging performance
counters in PerfMon to see how well BizTalk is performaing. (Orchestrations
Completed/second, messages processed/second etc.).
You can also create a test program in C#, or another language, if you want
to test more compex scenarios, such as HTTP adapters, MQSeries etc.
Post again if you need more advice,
Regards,
Alan
"Rick" wrote:
> Hi All,
>
> I am relatively new to BizTalk. We are working on a
> project where the first step is to figure out whether
> Biztalk caters to our needs.
>
> I have gone through the performance white paper. But I am
> not quite sure about how they measured the through put. We
> want to run our own tests and see BizTalk's performance.
> Please advise me how to calculate through put. We might
> try to simulate the tests done in Performance Paper.
>
> Thank you,
> Rick.
>
| |
|
| Hi Alan,
Thanks for your reply.
I have a file adapter with Orchestration. As suggested, I
looked at PERFMON too. But the problem is how to capture
end to end time consumed in addition to throughput?
I created simple scripts to simulate file drop. I am
working on scripts to capture data from WMI. But not quite
sure what data/events I should catch. We want to run
multiple tests altering performance parameters. Please
guide me.
Regards,
Rick
>-----Original Message-----
>Hi Rick,
>
>I have done a bit of basic performance testing on my
current project.
>
>The easyest way to test throughput is to use the file
adapter with an
>orchestration, and drop a bunch of say 1,000 xml files in
the recive folder.
>You can then use the XLANG/s Orchestrations and BizTalk
messaging performance
>counters in PerfMon to see how well BizTalk is
performaing. (Orchestrations
>Completed/second, messages processed/second etc.).
>
>You can also create a test program in C#, or another
language, if you want
>to test more compex scenarios, such as HTTP adapters,
MQSeries etc.
>
>Post again if you need more advice,
>
>Regards,
>
>Alan
>
>
>"Rick" wrote:
>
am[vbcol=seagreen]
We[vbcol=seagreen]
performance.[vbcol=seagreen]
>.
>
| |
|
| Have you considered using BAM since it can give you duration of process for
each message?
"Rick" wrote:
> Hi Alan,
>
> Thanks for your reply.
>
> I have a file adapter with Orchestration. As suggested, I
> looked at PERFMON too. But the problem is how to capture
> end to end time consumed in addition to throughput?
>
> I created simple scripts to simulate file drop. I am
> working on scripts to capture data from WMI. But not quite
> sure what data/events I should catch. We want to run
> multiple tests altering performance parameters. Please
> guide me.
>
> Regards,
> Rick
>
>
> current project.
> adapter with an
> the recive folder.
> messaging performance
> performaing. (Orchestrations
> language, if you want
> MQSeries etc.
> am
> We
> performance.
>
| |
|
| Alex,
I dont have MS office on my production machine. Is there
any alternative way to do this?
Thanks
>-----Original Message-----
>Have you considered using BAM since it can give you
duration of process for
>each message?
>
>
>"Rick" wrote:
>
I[vbcol=seagreen]
capture[vbcol=seagreen]
quite[vbcol=seagreen]
in[vbcol=seagreen]
BizTalk[vbcol=seagreen]
whether[vbcol=seagreen]
I[vbcol=seagreen]
put.[vbcol=seagreen]
might[vbcol=seagreen]
>.
>
| |
|
| Rick,
You only need to have MS office (Excel) installed on the client side.
It would be great to have web-based tools to monitor BAM activities, but I'm
not aware of anything similar available right now
AlexS.
"Rick" wrote:
> Alex,
>
> I dont have MS office on my production machine. Is there
> any alternative way to do this?
>
> Thanks
>
>
> duration of process for
> I
> capture
> quite
> in
> BizTalk
> whether
> I
> put.
> might
>
| |
| Matt Hall 2004-10-28, 7:48 am |
| We got around presicely this issue by creating a c# utility that inserted a
timestamp in a header of the message prior to submitting to BizTalk. Then,
depending on the pattern we were implementing, either the same c# utility or
an ASPX page was used as the endpoint that retrieved this timestamp and
calculated the latency/throughput. The results were logged to a SQL Database
and reports were created.
We have also created a pipeline component that is used to measure the timing
of the message through the infrastructure. I guess depending on the pattern
you are testing you could always implement something within a custom pipeline
to achieve the results you are after.
Also, as mentioned you could just utilise BAM.
I guess it all depends on your requirements, how in-depth, whether you wish
to extend, if you need to resuse at a later date in other scenarios, what
software is available etc, time constraints etc and then determine the best
method for you.
Matt Hall
http://www.bloglines.com/blog/matthall
"Rick" wrote:
> Alex,
>
> I dont have MS office on my production machine. Is there
> any alternative way to do this?
>
> Thanks
>
>
> duration of process for
> I
> capture
> quite
> in
> BizTalk
> whether
> I
> put.
> might
>
|
|
|
|
|