M Parker

We have two web servers: one is configured with gzip compression while the other one is not. I'd expect that when we run the load test against the web server w/gzip compression, the response time would be faster. But our testing shows both servers have similar response time. Any idea why



Re: Visual Studio Team System - Web and Load Testing Load test response time identical with or w/o gzip compression

Dennis Stone - MSFT

Does the webtest you're running against the two servers have the Accept-Encoding specified with gzip compression




Re: Visual Studio Team System - Web and Load Testing Load test response time identical with or w/o gzip compression

M Parker

I ran the webtest, reviewed the request headers from the results, and I could see "Accept-Encoding : gzip". What else should I check



Re: Visual Studio Team System - Web and Load Testing Load test response time identical with or w/o gzip compression

Dennis Stone - MSFT

Are you testing servers on an intranet Do you see the same results when you set the connection settings in your loadtest to use only very slow connections, such as 28.8




Re: Visual Studio Team System - Web and Load Testing Load test response time identical with or w/o gzip compression

M Parker

Yes, I am testing servers on an Intranet and yes, I am simulating the modem 28.8 connection.



Re: Visual Studio Team System - Web and Load Testing Load test response time identical with or w/o gzip compression

M Parker

Regardless of what connection speeds I use, the response time appears to be very similar. I know our web server is configured properly. When I ran the test using Fiddler, I could see the content length of most pages on the web server with gzip was about 1/4 the size of the one w/o.



Re: Visual Studio Team System - Web and Load Testing Load test response time identical with or w/o gzip compression

Dennis Stone - MSFT

Via team system are you seeing the same, or are you seeing the content-lengths being the same from both servers




Re: Visual Studio Team System - Web and Load Testing Load test response time identical with or w/o gzip compression

M Parker

The content-lengths displayed on the test results show the same results on both server via team system. I understand there is a bug on team system that it currently shows only the uncompressed content lengths. I am ok with that. I just want to be able to capture the response time correctly. Right now, the response times are identical on both servers. That's what puzzles me.



Re: Visual Studio Team System - Web and Load Testing Load test response time identical with or w/o gzip compression

Ed Glas - MSFT

The network emulation code uses the content-length to determine how much latency to add to a response. My guess is this same bug is tripping up the the bandwidth simulation.

Ed.






Re: Visual Studio Team System - Web and Load Testing Load test response time identical with or w/o gzip compression

Alan N

Was this issue reproduced and/or fixed for VSTS 2008 RTM
There is talk about experimenting with HTTP compression, and I've been asked whether VSTS can test it.

Thanks,
~Alan




Re: Visual Studio Team System - Web and Load Testing Load test response time identical with or w/o gzip compression

Alan N

Would still really like to know...anyone

~Alan