How IO wait times are calculated ? (4341583)

1.How does Spotlight gather and calculate the IO wait times?

ANSWER:
The figures are derived from the sys.dm_io_virtual_file_stats Dynamic Management View.
This DMV returns a table with the columns io_stall_read_ms and io_stall_write_ms amongst others. (See: http://technet.microsoft.com/en-us/library/ms190326.aspx)
A rate is calculated for these figures (by taking the difference between two samples and dividing by the time interval between the samples) and this is what is displayed in the grid.
Here's a page describing what appears to be a similar discrepancy http://sqlblog.com/blogs/michael_zilberstein/archive/2012/11/28/46419.aspx
So it would appear that SQL Server's view of what is happening can be quite different from the OS's view.

2.  What sample times Spotlight uses to collect the IO wait times?   Is it based on the background schedules (like "IO Stall") that we can configure? 

ANSWER:
In general, when we are on a Home Page or drilldown then it refreshes every 15 seconds, otherwise it is the background rate as it appears in the Scheduling dialog.  For the IO Wait times (Max I/O Wait control) it is the IO Stall collection, which by default is every 1 minute.

3. The MAX I/O Wait on the home screen, the balloon help states it is based on a specific number of background collections.  Is this configurable?

ANSWER:
So Max I/O Wait on the Home Page is 15 seconds if the page is open, otherwise 1 minute.
The alarm threshold uses an average over 5 collections which means that it varies depending on whether you have the home page open or not.  The user cannot control the number of collections the average is calculated over.  However the control shows the maximum value of the latest collection so it doesn't align very well with the alarm.
As a result of this we are investigating a better way of displaying this to make them consistent.

4. When drilling down into the home screen's MAX I/O Wait and select  I/O by File cannot see any wait numbers close to what is displayed on the home page.

ANSWER:
They are populated by different collections so the result could vary quite a lot.  For the drilldown I/O by File it is Virtual File Statistics.