"Peak IOPS over IO rate ranges"

I did an assessment recently and registered 6 virtual terminal servers in the appliance. The reason for this is that the customer experience a lot of performance issue on their virtual terminal servers. Max. 10 terminal server sessions on a terminal server with medium heavy users.

Now I see in the report "Peak IOPS over IO rate ranges" very high peak iops. 13000 IOPS and more. 3 terminal servers run on local storage and the spindles never deliver that much IOPS.

My Question:

how are the PEAK IOPS the this report "Peak IOPS over IO rate ranges" measured?

regards,

Edwin Voermans