Hi Tjark,

The question “Is my systems disk (subsystem) the bottleneck?” can be answered by looking at the Disk Queue Length. Queue length refers to the number of outstanding IO requests that are in the queue to be sent to the disk. This is measured as an absolute number of requests. The average number of seconds it takes to get a response from the disk. This is the total latency. 

When running Mendix (Studio Pro), you can measure the Disk Queue Length:

1. Press Windows key + R to open Run.
2. Enter resmon.exe and press Enter.
3. In the Resource Monitor window, expand sections for Disk, Network, and Memory.

In the Overview Tab, you can see the combination of the 4 main bottlenecks. Compile and run Mendix to get a feeling if the disks slow you down, or the other 3. You can also relate the resources used to the various other programs that are running in parallel.

If you go to the disk tab and compile, you will get the information on the Disk Queue Length

In the example above, I just re-compiled after changing a page.

In the next screenshot, I sorted on Read, and compiled from scratch:

Again you can see in the Queue Length Graph that there is some action:

The rule of thumb is that the Disk Queue Length of 2 or more is an indication of a potential issue. Not, a single spike, but for a longer period. In my system, you see the scale on the right: 0,05 so, I know there is no issue at all.

If you have a high value, and you implement RAID0, your value should drop by 50% of that value after implementation, as you have more disks to write, hence more Disk bandwidth.

Go Make It.

So, I guess a change to RAID0 would only lead to minor advantages.