Are there any known and sound benchmarks comparing a standard Windows Server 2008 with a Server Core installation? I guess a realistic workload like SQL Server or a web application running under IIS would be important.
-
I don't think Core vs server was done to provide a performance gain, although it likely would in a large enough environment. I think the driver is to reduce the vulnerability footprint– uSlackrApr 13, 2012 at 13:47
-
@uSlackr, I think you are right but one might argue that less memory usage and less CPU cache churn might improve performance. Also, less periodic background activity.– usrApr 13, 2012 at 13:49
-
The RAM footprint is certainly smaller, but I don't think it will have much impact on CPU cycles, to be honest.– MDMarraApr 13, 2012 at 14:05
-
@MDMarra that might be true, but it would be interesting to know for sure!– usrApr 13, 2012 at 14:05
Add a comment
|
1 Answer
Just benchmark these services like you normally would with command-line tools that work on both like sqlio.