SYSTEM AND METHOD OF MIRRORED RAID ARRAY WRITE MANAGEMENT

• US 20080059699A1
• Filed: 09/06/2006
• Published: 03/06/2008
• Est. Priority Date: 09/06/2006
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Kubo, Robert A., Pinson, Jeremy M.
• Granted Patent: US 7,555,599 B2
• Time in Patent Office: Days
• US Class Current: 711/114
• CPC Class Codes: G06F 11/2087 with a common controller
• CPC Class Codes: G06F 3/0613 in relation to throughput
• CPC Class Codes: G06F 3/0656 Data buffering arrangements
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD

Abstract

The write operations to the storage devices are managed so that the write operations that would force a storage device to reposition its read/write head outside its read optimization boundary are queued until a threshold is reached. Upon reaching the write queue threshold, the write operations are sent to the storage devices in a burst mode that limits the impact of storage device seek operations that force the device read/write head out of position relative to the data that is assigned as it'"'"'s primary read region, keeping the read/write head in it'"'"'s primary locality to optimize the performance of read operations.

Claims

1. A method of mirrored RAID array write management, comprising:

receiving a write request to a mirrored RAID array;

queuing the write request to a write queue;

determining whether the write queue reached a selected threshold;

if the write queue has reached the selected threshold,

ordering a plurality of write requests in the write queue sequentially to map to logical block addresses of a hard disk drive; and

transferring the plurality of write requests to the hard disk drive one at a time in a burst of transactions.

2. The method of claim 1, further including:

servicing one or more read operations that are hits to the write cache from the write cache.

3. The method of claim 1, wherein the write request is received from a RAID adapter writer buffer.

4. The method of claim 1, wherein the selected threshold is determined based on an amount of data in the write queue.

5. The method of claim 1, wherein the selected threshold is determined based on latency.

6. The method of claim 1, wherein the selected threshold is determined based on a number of I/Os pending.

7. The method of claim 1, wherein the steps of claim 1 are performed for a plurality of write requests to a region of a hard disk drive that is outside an optimal area.

8. A method for mirrored RAID array write management, comprising:

determining if a write queue reached a threshold;

if the write queue has reached a threshold, sending queued write requests in the write queue to a hard disk drive;

determining whether a write request is for a data location in a read optimized region;

if the write request is for a data location in a read optimized region, sending the write request to a hard disk drive; and

if the write request is not for a data location in a read optimized region, adding the write request to the write queue.

9. The method of claim 8, wherein the threshold is a predetermined number of write requests.

10. The method of claim 8, wherein the threshold is a predetermined time period.

11. The method of claim 8, further including:

if the write request is not for a data location in read optimized region, determining whether a write queue exists; and

if the write queue does not exist, initializing a write queue and starting write queue age timer.

12. The method of claim 11, further including:

placing targeted data location of the write request to a pending status.

13. The method of claim 8, further including:

placing targeted data location of the write request to a pending status.

14. A system for mirrored RAID array write management, comprising:

means for receiving a write request to a mirrored RAID array;

a write queue;

means for queuing the write request to the write queue;

means for determining whether the write queue reached a selected threshold;

means for ordering a plurality of write requests in the write queue sequentially to map to logical block addresses of a hard disk drive; and

means for transferring the plurality of write requests to the hard disk drive one at a time in a burst of transactions, if the write queue has reached the selected threshold.

15. The system of claim 14, wherein the means for queuing queues the write request if the write request is not to a data location in read optimized region.

16. The system of claim 14, wherein the selected threshold includes a predetermined number of write requests.

17. The system of claim 14, wherein the selected threshold includes a predetermined time period.