Application profiling in a data storage array

An apparatus comprising:

• a storage array of independent data storage devices that form a fast pool and a slow pool of said devices;
• and
• a controller adapted to, responsive to receipt of a first hint that a selected application is about to be executed a first time, makes a return on investment (ROI) determination with regard to a requested distributed data set which forms a part of the selected application and which is stored across a first plurality of the devices in the slow pool, said ROI determination indicating that a cost of migrating the distributed data set to a second plurality of said devices in the fast pool will be outweighed by an overall improved data transfer capacity of the storage array over a predetermined minimum payback period of time;
• wherein
• the controller is further adapted to, responsive to the ROI determination and to receipt of a second hint that the selected application is about to be executed a second time subsequent to the first time, migrates the distributed data set to the second plurality of said devices in the fast pool.

A method comprising:

• arranging a distributed data storage array of data storage devices to respectively form a fast pool and a slow pool of said devices;
• receiving a hint that a selected application is about to be executed;
• and
• migrating a selected distributed data set stored across a first plurality of said devices in the slow pool to a second plurality of said devices in the fast pool responsive to said hint and responsive to a return on investment (ROI) determination that a cost associated with said migration is justified by an overall improved data transfer capacity of the storage array over a predetermined minimum payback period of time.

A method comprising:

• arranging a distributed data storage array of data storage devices to respectively form a fast pool and a slow pool of said devices;
• storing a selected distributed data set across a first plurality of said devices in the slow pool;
• receiving a first hint that a selected application stored in the array is about to be executed a first time, the selected application including the selected distributed data set;
• recording migrations of data to the fast pool over an activity period immediately following receipt of the first hint in an activity log in a memory space, said migrations including a first migration of the selected distributed data set to said fast pool based on a return on investment (ROI) determination;
• returning the selected distributed data set back to the slow pool at a conclusion of the first execution of the selected application;
• subsequently receiving a second hint that the selected application is about to be executed a second time;
• and
• performing a subsequent, second migration of the distributed data set from the slow pool to the fast pool prior to a user-initiated access request for said data set during said second accessing of the selected operation responsive to the activity log and the second hint.