Optimizing thin provisioning in a data storage system through selective use of multiple grain sizes

A data storage system, comprising:

• a pre-cache;
• a management node that manages a plurality of storage devices across which a data storage array is distributed, wherein the management node is configured to perform:
• in response to the data storage system receiving, from a host, a write request specifying a logical address and write data, sending acknowledgement of the write request to the host and buffering the write request as one of a plurality of write requests concurrently buffered in the pre-cache without provisioning, in the data storage array, a physical extent corresponding to the logical address;
• analyzing, prior to provisioning in the data storage array any physical extent for the plurality of write requests, at least sizes of the plurality of write requests buffered in the pre-cache;
• thin provisioning physical extents in the data storage array on an as-needed basis based on the analyzing, wherein the thin provisioning includes:
• in response to the analyzing identifying a first pattern of write requests in the plurality of write requests, the management node provisioning, in the data storage array, a first physical extent having a smaller grain size and destaging write data of the plurality of write requests from the pre-cache to the first physical extent;
• in response to the analyzing identifying a second pattern of write requests in the plurality of write requests, the management node provisioning, in the data storage array, a second physical extent having a larger grain size and destaging write data of the plurality of write requests from the pre-cache to the second physical extent;
• and
• repeating the buffering, analyzing and thin provisioning for each subsequent plurality of write requests received from the host and buffered in the pre-cache.

A program product for thin provisioning physical storage in a data storage system including a pre-cache and a plurality of storage devices across which a data storage array is distributed, the program product comprising:

• a data storage device;
• and
• program code stored within the data storage device, wherein the program code, when executed by a processor of a management node of the data storage system, causes the management node to perform:
• in response to the data storage system receiving, from a host, a write request specifying a logical address and write data, sending an acknowledgement of the write request to the host and buffering the write request as one of a plurality of write requests concurrently buffered in the pre-cache without provisioning, in the data storage array, a physical extent corresponding to the logical address;
• analyzing, prior to provisioning in the data storage array any physical extent for the plurality of write requests, at least sizes of the plurality of write requests concurrently buffered in the pre-cache;
• thin provisioning physical extents in the data storage array on an as-needed basis based on the analyzing, wherein the thin provisioning includes:
• in response to the analyzing identifying a first pattern of write requests in the plurality of write requests, provisioning, in the data storage array, a first physical extent having a smaller grain size and destaging write data of the plurality of write requests from the pre-cache to the first physical extent;
• in response to the analyzing identifying a second pattern of write requests in the plurality of write requests, provisioning, in the data storage array, a second physical extent having a larger grain size and destaging write data of the plurality of write requests from the pre-cache to the second physical extent;
• and
• repeating the buffering, analyzing and thin provisioning for each subsequent plurality of write requests received from the host and buffered in the pre-cache.