Need help with StresStimulus? Start here.
Browser cache settings for a Silverlight application
Unregistered User Posts: 514
6/6/2013
|
This question was originally posted by Shasha Han in this post. It was reposted here to create a separate thread, to make it easier for the Community to find it.
Our system is a silverlight application. When it run in Explore, it depends on the explore's setting to read cache or not. It leads to different time when you open system at the first time and later(if using cache, explore will save a lot of time of preparing silverlight app since it has already downloaded bofore). I'm wondering how StresStimulus treat cache. Can I configurate it before I running of it? Or it ignore cache when load test running , for each duration , for each person , it will clear cache before next duration?
|
|
0
link
|
Vadim @StresStimulus Administrator Posts: 583
6/7/2013
|
StresStimulus has capabilities to emulate browser cache. To use this feature go to Multi Test Cases , select your test case and enable the cache control property. This will tell StresStimulus to follow the same caching rules that are implemented in web browsers. These rules include multiple factors including caching response headers and emulated browser cache content ( empty vs. primed) for each resource.
Once cache control is enabled, you can set the percent of new VUs vs. returning VUs. New VUs will simulate an empty browser cache on the first iteration and primed cache on the subsequent, while returning VUs will simulate primed cache in all iterations. You can also configure % VUs restarting browsers before starting new iterations. For these VUs, browser sessions will not persist across test iterations.
|
|
0
link
|
Shasha Han Posts: 8
7/11/2013
|
Hi, Vadim We have bought your StresStimulus for the first month. As a qa person, I'm the first one who to use your product. I have tried to use it for two days, I have a few questions, they are still about cache.
As your indication, I have known about the the New VUs percent and VUs restarting browsers% but I'm wondering if my way is correct. what I want is all VU can use cache and session and I set New VUs% as 1, VUs restarting browsers% as 0. I think my configuration is nearly totally to use cache for every VU. is that correct?
The problem is our system is a silverlight system. It will download Silverlight package including caching data when you open the web at the first time. From the test result, the request is:
/ClientBin/UniteamTransportSystems.xap
|
Content Type is :application/x-silverlight-app
The Response time is 3.103, and the Network Time is 3.055.
What I understand is the download time is 3.055, is that correct? The question is I have set Cache Control Part as I mentioned above, I thought the Net workTime should be nearly zero since I have download it at the first time. But what I get is kind of same as the situation that I set CacheConrol as disabled b/c the time is still about 3 second when I didn't use Cache(Cache Control is disable mean I don't use cache, right?) So I think I must do something wrong.... Would you pls point me out?
And another question is, what's meaning of TTFF(s), I noticed that it plus Network time equal Response time, could you explain more to me?
I'm still a new person on load test area, my question may be not professional, thank you in advance
shasha
|
|
0
link
|
Vadim @StresStimulus Administrator Posts: 583
7/12/2013
|
Hi Shasha, Your configuration provides that 99% of VUs are returned users, 1% of VUs are new users, and all VUs keep browser open on the second and subsequent iterations. The browser Cache Control rules are explained in paragraph 1 of this post http://www.stresstimulus.com/blog/post/new-features-in-v1-2-final-release-part-1 When a returned user requests the Silverlight component, StresStimulus will either send a conditional request or will not even send a request (it depends on the Silverlight component response caching headers). In the first case, your server probably should return 304, which is essentially an empty response. If this is not happening, you have to troubleshoot.
As far as TTFF, you probably meant TTFB - time to first byte http://en.wikipedia.org/wiki/Time_To_First_Byte It is very close to server time. For static files (like Silverlight component ) it should be close to zero. You're correct that TTFB + Network Time = Response Time
Cheers, -Vadim
|
|
0
link
|
Copyright © 2024 Stimulus Technology