[Jonathan Eason]

Hi Eric,

The latency limit is based on the type of work flow, either File Based (i.e. Frame Based) or streaming. For file based the calculation is simple. Just answer how many frames you want per second. At 24 fps you could not go over 41ms. Keep in mind this is a total value from request to delivery for the application and includes the entire i/o stack (i.e. application, storage, hardware drivers, operating system). Most applications will buffer and hopefully perform threaded requests which will keep you within acceptable limits. I would just benchmark from the client side to any shared volumes with single threaded i/o to establish the baseline. Just be sure that your benchmark tool can emulate your editing software otherwise your benchmarks are worthless.

Cheers,
Jonathan

“A Day of life spent learning nothing is a day of life wasted.” ~ Anonymous

[Jonathan Eason]

Sounds like you will need to re-tune the system. I would start with narrowing down the i/o paths to single routes and perform a baseline benchmark with ONE tool. Windows Server has built-in Performance Monitoring or there is Process Monitor that you can download via the Windows Sysinternals website. I prefer Process Monitor becuase it is easier to use. If you have both environments running it should be relatively easy to isolate areas of latency within your applications.