I think you’re right that there’s potentially going to be an impact of moving data around dynamically and that’s why “FAST for blocks” rather than “FAST for LUNs” isn’t going to be available until 2010 – gives EMC plenty of time to get it to work. I think it’s already been mentioned by EMC that their thin provisioning implementation is a precursor to being able to place/read those blocks from disparate hardware devices.

At this stage, as FAST for Blocks is essentially slideware, we should all be expecting to see EMC providing performance figures to back up how well FAST will perform. On current form, that’s unlikely.

Chris

I think you might want to consider the impact of moving data from one tier to another every few minutes. Do you really want all that data on the backend moving around all the time? This seems like a performance killer for sure.

Compellent’s Data Progression evaluates data every 24 hours moving that data in minutes or seconds but only once a day.

WAC

Isn’t Compellent’s technology based over a longer timescale? I thought the Compellent solution took data over a longer time to do the determination of where data should sit. A bit like EMC’s Optimiser.

Chris

FAST, is it not the smae technology EMC was criticising few years back, and I think they stole the patent from Compellent or 3Par, which I believe is preparing to get into a legal battle. However, given EMC is clout and money power they are unlikley to make much dent.

Ravi