Re: Has performance improved?

Hi Gary,

Thanks for your prompt response. I really appreciate that

1. Isn't total page load time and response time of individual http requests interlinked? I mean, better response time of individual requests lead to better page load time? Coming to your question, the KPI is the total page load time

2. Yes, builds A and B are functionally equivalent. Build A contains code that is currently deployed in production and build B has performance tuned code. The prob in B is that it has so many additional ".ashx" requests & even though the ".aspx" calls which are common in both A & B has improved marginally in B for a page, am not sure if those additional ".ashx" call, will increase the overall page load time

3. I agree that the TPS can be used as an indicator to measure performance improvement. But, if the TPS is same(or very close) for both the runs, does it mean the performance hasn't improved much?

Regards,

Chandiran

On Dec 3, 2014 12:25 AM, Gary Mulder <[email protected]> wrote:

On 2 December 2014 at 17:30, Zulu Zulu <[email protected] > wrote:

Hi All,

I spent three months in learning & implementing this wonderful tool in my organization.

I recorded the scripts for two different builds (A and B) using TCPProxy, modified according to my requirements, ran it and collected the two different logs. The dev team has made lots of changes in the application to improve the performance of our application (which is build B). Now, the dev along with the higher management wants to know whether the application performance has actually improved or not. I proposed the following method which got rejected

* Add the average mean time of all requests for each page for both builds A & B and compare them. They rejected this outright saying the builds (A & B) may vary with the number of requests in each page. And since I do a sum of the average mean time of all the requests in a page, the load time will definitely be higher for the page having higher requests. (These requests are actually resource files which in real browser will be loaded in parallel).

I'm now in a position where I really could not explain to them whether the performance has actually improved or not.

Would really appreciate if someone can guide me on this

You need your stakeholders to define their KPIs (Key Performance Indicators). Is it total page load time or response time for individual HTTP calls?

Are A and B functionally equivalent? Why can't you run recording A against B? If you need to compare recording A and B, and B is resulting in faster page load times (i.e. TPS for B has increased), you could reduce the number of threads until you get approximately the same TPS for recording B as you did for recording A. Then look at the response times for calls that are common for both recordings and see if response times have improved for them.

Hope that helps...

Gary

------------------------------------------------------------------------------
Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
from Actuate! Instantly Supercharge Your Business Reports and Dashboards
with Interactivity, Sharing, Native Excel Exports, App Integration & more
Get technology previously reserved for billion-dollar corporations, FREE
http://pubads.g.doubleclick.net/gampad/clk?id=164703151&iu=/4140/ostg.clktrk

_______________________________________________
grinder-use mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/grinder-use