Hybrid storage subsystem with mixed placement of file contents

摘要

A storage subsystem combining solid state drive (SSD) and hard disk drive (HDD) technologies provides low access latency and low complexity. Separate free lists are maintained for the SSD and the HDD and blocks of file system data are stored uniquely on either the SSD or the HDD. When a read access is made to the subsystem, if the data is present on the SSD, the data is returned, but if the block is present on the HDD, it is migrated to the SSD and the block on the HDD is returned to the HDD free list. On a write access, if the block is present in the either the SSD or HDD, the block is overwritten, but if the block is not present in the subsystem, the block is written to the HDD.

主权项

1. A computer-performed method of managing a storage subsystem including a lower-latency block storage device and a higher-latency block storage device, the method comprising:
receiving a file read access operation request implicating multiple blocks of a single file corresponding to the file read access operation, wherein if the file implicated by the file access operation request is present in the storage subsystem, individual ones of the multiple blocks are stored exclusively either on the lower-latency block storage device or the higher-latency block storage device, wherein a first set of the multiple blocks corresponding to a first portion of the file are stored on the lower-latency block storage device, and wherein a second set of the multiple blocks corresponding to a second portion of the file and exclusive of the first portion of the file are stored on the higher-latency block storage device simultaneously with the first set of the multiple blocks being stored on the lower-latency block storage device; specifying a next block in the file implicated by the request as a given block; determining whether the given block is present in the lower-latency block storage device; responsive to determining that the given block is present in the lower-latency block storage device, accessing the given block at the lower-latency block storage device; responsive to determining that the given block is not present in the lower-latency block storage device, accessing the given block at the higher-latency block storage device to copy the given block to the lower-latency block storage device and remove the given block from the higher-latency block storage device; repeating the specifying, determining and accessing for following blocks implicated by the file access operation request until a last block implicated by the file read access operation request has been accessed; receiving a file write access operation request implicating other multiple blocks; specifying a next block in another file implicated by the write request as another given block; determining whether the another given block is present in either of the lower-latency block storage device or the higher-latency block storage device; responsive to determining that the another given block is not present in the lower-latency block storage device nor the higher-latency block storage device, allocating a new block on the higher-latency block storage device; and responsive to allocating the new block on the higher-latency block storage device, writing the data provided with the write operation only to the new block, wherein if the another given block is not present in either of the lower-latency block storage device or the higher-latency block storage device, the data provided with the write operation is always written exclusively to the higher-latency block storage device, whereby the another given block is stored exclusively on the higher-latency block storage device.