Storage systems using write off-loading

• US 8,074,014 B2
• Filed: 03/31/2008
• Issued: 12/06/2011
• Est. Priority Date: 03/31/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Narayanan, Dushyanth, Donnelly, Austin, Rowstron, Antony, Thereska, Eno, Elnikety, Sameh
• Primary Examiner(s): DUDEK JR, EDWARD J
• Application Number:
  US12/059,695
• Publication Number:
  US 20090249001A1
• Time in Patent Office: 1,345 Days
• Field of Search: 713/324, 711/154
• US Class Current: 711/112
• CPC Class Codes: G06F 1/3203 Power management, i.e. event-based initiation of a power-saving mode
• CPC Class Codes: G06F 1/3268 Power saving in hard disk drive
• CPC Class Codes: G06F 11/1441 Resetting or repowering
• CPC Class Codes: G06F 11/1471 involving logging of persistent data for recovery
• CPC Class Codes: G06F 11/3433 for load management allocation of a server based on load conditions G06F9/505; load rebalancing G06F9/5083; redistributing the load in a network by a load balancer H04L67/1029
• CPC Class Codes: G06F 11/3485 for I/O devices
• CPC Class Codes: G06F 2201/82 Solving problems relating to consistency
• CPC Class Codes: G06F 3/0625 Power saving in storage systems
• CPC Class Codes: G06F 3/0634 by changing the state or mode of one or more devices
• CPC Class Codes: G06F 3/0689 Disk arrays, e.g. RAID, JBOD
• CPC Class Codes: Y02D 10/00 Energy efficient computing, e.g. low power processors, power management or thermal management

Abstract

Improved storage systems which use write off-loading are described. When a request to store some data in a particular storage location is received, if the particular storage location is unavailable, the data is stored in an alternative location. In an embodiment, the particular storage location may be unavailable because it is powered down or because it is overloaded. The data stored in the alternative location may be subsequently recovered and written to the particular storage location once it becomes available.

Claims

1. A method of storing data comprising:

receiving a request from a client to store data on a first storage device;

determining if defined criteria are met;

storing a most recent version of the data, if the defined criteria are met, on a second storage device and not on the first storage device by sending a write request to the second storage device, the second storage device comprising a circular log;

maintaining an identity of the second storage device amongst a set of identified storage devices in a hard state data structure;

maintaining logical block numbers for circular log blocks storing the data and a version number of the most recent version of the data in a soft state redirect data structure; and

maintaining locations of older versions of the data in the circular log in a soft state garbage data structure until the older versions are invalidated.

2. A method according to claim 1, wherein determining if defined criteria are met comprises:

determining if the first storage device is powered down.

3. A method according to claim 2, wherein determining if defined criteria are met further comprises:

determining if an older version of at least a portion of the data is stored on the second storage device.

4. A method according to claim 1, wherein determining if defined criteria are met comprises at least one of:

determining if the first storage device is overloaded; and

determining if the first storage location has a higher load than an alternative storage location.

5. A method according to claim 1, wherein storing the data on the second storage device comprises:

selecting a second storage device from the set of identified storage devices based on load metrics for each of the set of identified storage devices;

sending a request to store the data to the second storage devices; and

receiving an acknowledgement from the second storage devices confirming that the data has been stored.

6. A method according to claim 5, further comprising:

selecting at least one other storage devices from the set of identified storage devices based on load metrics for each of the set of identified storage devices; and

sending the request to store the data to each of the selected storage devices.

7. A method according to claim 1, further comprising:

recovering the data from the second storage devices and storing it on the first storage device.

8. A method according to claim 7, further comprising:

sending, to the second storage device, a request to invalidate the most recent version of the data and any older versions of the data.

9. A method according to claim 1, further comprising:

determining if the first storage device is idle; and

powering down the first storage device if the first storage device is idle.

10. A method according to claim 9, further comprising:

powering up the first storage device on receipt of a request from a client to read data stored on the first storage device.

11. A method according to claim 1, further comprising employing the hard state data structure, the soft state redirect data structure, and the soft state garbage data structure to recover the data in an event of a failure.

12. A storage system comprising:

a plurality of storage devices; and

a storage manager, associated with a first storage device from the plurality of storage devices;

a logger module associated with a second storage device, wherein the logger module is arranged to;

receive a write request from the storage manager when the first storage device is unavailable, the write request comprising data and metadata;

store the data and the metadata in the second storage device, the second storage device comprising a circular log including one or more blocks used to store the data and metadata;

maintain a hard state data structure configured to store multiple versions of the same block, at least one version being a most recently written version and at least one version being a stale version; and

maintain a soft state data structure that stores a list of most recently written versions for a plurality of different blocks stored for the storage manager, but does not store details relating to stale versions of the plurality of different blocks.

13. A storage system according to claim 12, wherein the first storage device comprises a volume of disks and wherein the second storage device comprises a partition on a second volume of disks.

14. A storage system according to claim 12, wherein the storage manager is arranged to:

intercept a write request to the first storage device; and

forward the write request to the second storage device.

15. A storage system according to claim 14, wherein forwarding the write request to the second storage device when the first storage device is unavailable comprises at least one of:

forwarding the write request to the second storage device if the first storage device is powered down;

or

forwarding the write request to the second storage device if the first storage device is overloaded.

16. A storage system according to claim 12, wherein the soft state data structure enables the logger module to find the most recent version of a block in the circular log.

17. A storage system according to claim 12, wherein the soft state data structure is recreated from the hard state data structure in the event of a failure.

18. One or more tangible device-readable media with device-executable instructions configured to perform acts comprising:

receiving a write request to a storage location, the write request comprising data;

storing a most recent version of the data in the storage location, the storage location comprising a circular log;

maintaining an identity of the storage location amongst a set of identified storage locations in a hard state data structure;

maintaining logical block numbers for circular log blocks storing the data and a version number of the most recent version of the data in a soft state redirect data structure; and

maintaining locations of older versions of the data in the circular log in a soft state garbage data structure until the older versions are invalidated.

19. The one or more tangible device-readable media of claim 18, wherein the device-executable instructions are further configured to perform acts comprising:

determining that a first storage location, for which the write request is directed, is unavailable;

if the first storage location is unavailable;

selecting the storage location as an alternative when the first storage location is unavailable; and

forwarding the write request to the storage location.

20. The one or more tangible device-readable media of claim 18, wherein the device-executable instructions are further configured to employ the hard state data structure, the soft state redirect data structure, and the soft state garbage data structure to recover the data in an event of a failure.