2

I am running an apache server and a JSP web application which I am accessing over LAN.
I want to test the performance of the web app(it's a third party...not written by me) to decide whether to run it under HTTP or HTTPS.

I am a beginner when it comes to all this performance testing and all. Can someone recommend me tools and give some beginner level info on this issue(a detailed blog post would also do).
Basically, I need to know how should I go about the testing and how can I decide whether to run it under HTTP or HTTPS?
Ofcourse HTTPS would be preferred but if the performance impact is significant then I will run it under HTTP. I want to know how the quantified impact on performance.

I know there are some similar questions out there(like this How much of a performance hit for https vs http for apache?), but the answers are mostly targeted towards quite expert administrators and I am not able to understand how to use the tools mentioned in those answers.

P.S. I am running everything under windows.

Community
  • 1
tumchaaditya
  • 121
  • 2
  • Capacity and planning questions aren't suited to Server Fault. If you want to test performance and find the best test tool - replicate the load of whatever you are doing. With regards to HTTP vs HTTPS - the bottleneck is generally in CPU time for the encryption/decryption. – Ben Lessani Sep 28 '12 at 12:14
  • @sonassi: IME, CPU time rarely significant - unless you use a very specific subset of the elliptical curve algorithms – symcbean Sep 28 '12 at 13:42
  • I can't comment on JSP - we host PHP E-Commerce stores (Magento). And during testing, the difference between HTTP and HTTPS is noticeable resulting in increased CPU load. – Ben Lessani Sep 28 '12 at 14:00

1 Answers1

3

In practice, unless you're running on an absolutely minimal system the performance difference is negligible. Nevertheless, performance shouldn't even be a concern. Unless you have a need to secure the connection it makes absolutely no sense to use HTTPS. Conversely, if you need to run HTTPS the performance difference compared to HTTP is totally irrelevant.

John Gardeniers
  • 27,458
  • 12
  • 55
  • 109