Title:

Disk-drive system having multiple power supplies and mirroring and method

Document Type and Number:

United States Patent 7447015

Abstract:

A system and method for an improved multiple hard-disk-drive data-storage enclosure. Some embodiments position drives in counter-rotating pairs, each simultaneously accessing half the data, such that seek-caused actuator rotational-acceleration vibration cause simultaneous canceling rotational torque. Some embodiments position the edge of a first drive (or drive pair) at an angle to the actuator midpoint of a nearby second drive (or drive pair), such that rotational-acceleration vibration from a seek-caused actuator rotation in the first drive does not cause a rotational movement into the second drive that affects the tracking or seek operation. Some further embodiments position drives in a herringbone pattern to redirect air flow in addition to reducing rotational-acceleration vibration interaction. Other embodiments include a printed wire circuit board mounted to reduce the rotational-acceleration vibration interaction.

Inventors:

Hall, Jonathan E. (Golden, CO, US)
Mccormick, Daniel M. (Superior, CO, US)
Wendel, Eric J. (Johnstown, CO, US)
Lemaire, Charles A. (Apple Valley, MN, US)

Plaque It!

Application Number:

11/537603

Publication Date:

11/04/2008

Filing Date:

09/29/2006

View Patent Images:

Images are available in PDF form when logged in. To view PDFs, Login or Create Account (Free!)

Export Citation:

Click for automatic bibliography generation

Assignee:

Atrato, Inc. (Westminster, CO, US)

Primary Class:

713/320

Other Classes:

323/282, 365/200, 361/685

International Classes:

H05K5/00

Field of Search:

365/189.09, 713/300, 439/62, 323/282, 713/1, 361/18, 360/97.02, 361/788, 439/64, 323/222, 713/320, 360/137, 365/185.23, 361/724-727, 361/796, 360/69, 361/679-687, 323/308, 360/55, 361/785, 365/200

US Patent References:

4433388	Longitudinal parity	February, 1984	Oosterbaan
4761785	Parity spreading to enhance storage access	August, 1988	Clark et al.
5435737	Removable memory modules	July, 1995	Haga et al.
6154361	Disk-drive chassis for reducing transmission of vibrations between disk-drive units of a disk-drive array	November, 2000	Anderson et al.
6288902	Modular data storage system for reducing mechanical shock and vibrations	September, 2001	Kim et al.
6292360	Packaging system for mass memory units having uniform or mixed form factors	September, 2001	Carteau
6388873	Disk drive including resilient securing system providing relative movement between side rails and head disk assembly to accommodate side rails engaging guide channels in a chassis	May, 2002	Brooks et al.
6487039	Disc-drive mounting method and apparatus to reduce noise	November, 2002	Bernett
6574721	Apparatus and method for providing simultaneous local and global addressing using software to distinguish between local and global addresses	June, 2003	Christenson et al.
6987674	Disk storage system with removable arrays of disk drives	January, 2006	El-Batal et al.
7345296	Nanotube transistor and rectifying devices	March, 2008	Tombler et al.	257/9
7391609	Disk-drive enclosure having laterally offset parallel drives to reduce vibration and method	June, 2008	Hall et al.	361/685
20040193791	Midplane-independent implementations of data storage system enclosures	September, 2004	Felton et al.
20070279790	ADJUSTABLE OUTPUT VOLTAGE REGULATOR FOR DISK DRIVE	December, 2007	Tanner	360/69

Primary Examiner:

Duong, Hung V.

Attorney, Agent or Firm:

Lemaire, Charles A.
Lemaire Patent Law Firm, P.L.L.C.

Parent Case Data:

CROSS-REFERENCES TO RELATED INVENTIONS

This is a divisional of U.S. patent application Ser. No. 11/027,777, filed Dec. 29, 2004 and titled “SYSTEM AND METHOD FOR MASS STORAGE USING MULTIPLE-HARD-DISK-DRIVE ENCLOSURE,” which claims benefit of U.S. Provisional Patent Application No. 60/580,987, filed Jun. 18, 2004 and titled “SYSTEM AND METHOD FOR REDUCED VIBRATION INTERACTION IN A MULTIPLE-HARD-DISK-DRIVE ENCLOSURE,” and of U.S. Provisional Patent Application No. 60/533,605, filed Dec. 29, 2003 and titled “SYSTEM AND METHOD FOR IMPROVED HARD-DISK-DRIVE DATA-STORAGE ENCLOSURE,” each of which is hereby incorporated by reference in its entirety. This application is also related to U.S. patent application Ser. No. 11/026,553, filed Dec. 29, 2004 and titled “SYSTEM AND METHOD FOR REDUCED VIBRATION INTERACTION IN A MULTIPLE-DISK-DRIVE ENCLOSURE,” which is incorporated herein by reference in its entirety.

This application is additionally related to:

U.S. patent application Ser. No. 11/537,600 filed on Sep. 29, 2006 and entitled “DISK-DRIVE ENCLOSURE HAVING FRONT-BACK ROWS OF SUBSTANTIALLY PARALLEL DRIVES AND METHOD”;
U.S. patent application Ser. No. 11/537,605 filed on Sep. 29, 2006 and entitled “DISK-DRIVE ENCLOSURE HAVING ROWS OF ALTERNATELY FACING PARALLEL DRIVES TO REDUCE VIBRATION AND METHOD”;
U.S. patent application Ser. No. 11/537,606 filed on Sep. 29, 2006 and entitled “DISK-DRIVE ENCLOSURE HAVING LATERALLY OFFSET PARALLEL DRIVES TO REDUCE VIBRATION AND METHOD”;
U.S. patent application Ser. No. 11/537,608 filed on Sep. 30, 2006 and entitled “DISK-DRIVE ENCLOSURE HAVING NON-PARALLEL DRIVES TO REDUCE VIBRATION AND METHOD”;
U.S. patent application Ser. No. 11/537,614 filed on Sep. 30, 2006 and entitled “DISK-DRIVE ENCLOSURE HAVING PAIR-WISE COUNTER-ROTATING DRIVES TO REDUCE VIBRATION AND METHOD”;
U.S. patent application Ser. No. 11/537,598 filed on Sep. 29, 2006 and entitled “DISK-DRIVE ENCLOSURE HAVING DRIVES IN A HERRINGBONE PATTERN TO IMPROVE AIRFLOW AND METHOD”;
U.S. patent application Ser. No. 11/537,607 filed on Sep. 30, 2006 and entitled “DISK-DRIVE SUPPORTING MASSIVELY PARALLEL VIDEO STREAMS AND METHOD”;
U.S. patent application Ser. No. 11/537,610 filed on Sep. 30, 2006 and entitled “DISK-DRIVE SYSTEMS WITH VARYING NUMBER OF SPARES FOR DIFFERENT EXPECTED LIFETIMES AND METHOD”;and
U.S. patent application Ser. No. 11/537,613 filed on Sep. 30, 2006 and entitled “PORUOS LIGHT-EMITTING DISPLAY WITH AIR FLOW THROUGH DISPLAY, ITS USE IN A DISK-DRIVE SYSTEM AND METHOD”; which are all hereby incorporated by reference in their entirety.

Claims:

What is claimed is:

1. A method comprising: providing a single drive enclosure holding a plurality of disk drives including a first, a second, a third, and a fourth drive positioned in a pattern of disk drives that extends generally from an air-inlet side of the enclosure to an air-outlet side of the enclosure; performing a first conversion of input DC power into a first regulated DC voltage source; performing a second conversion of input DC power into a second regulated DC voltage source; connecting the first regulated DC voltage source to the first drive and the second drive but not to the third drive or the fourth drive; connecting the second regulated DC voltage source to the third drive and the fourth drive but not to the first drive or the second drive; and writing a first portion of a set of data to the first drive and a second portion of the set of data to the second drive and a mirrored copy of the first portion of the set of data to the third drive and a mirrored copy of the second portion of the set of data to the fourth drive such that if either the first or the second conversion of input DC power to the respective first and second voltage source fails then both the first portion and the second portion of the set of data remain available from drives powered by the other non-failing voltage source.

2. The method of claim 1, further comprising: performing a first AC-to-DC conversion to generate a first source of input DC power and a second AC-to-DC conversion to generate a second source of input DC power both external to the enclosure; connecting the first source of input DC power from the first AC-to-DC conversion for the first regulated conversion of input DC power into the first regulated DC voltage; and connecting DC power from the second AC-to-DC conversion for the second regulated conversion of input DC power into the second regulated DC voltage.

3. The method of claim 1, further comprising: sequencing the connecting of power from the respective voltage sources to the respective disk drives such that the first disk drive is powered up before the second disk drive, and the third disk drive is powered up before the fourth disk drive.

4. The method of claim 1, further comprising positioning the disk drives such that the pattern of disk drives positions: the first drive and a plurality of other drives in a first row of drives that extends generally from an air-inlet side of the enclosure to an air-outlet side of the enclosure, the second drive and a plurality of other drives in a second row of drives that extends generally from an air-inlet side of the enclosure to an air-outlet side of the enclosure, the third drive and a plurality of other drives in a third row of drives that extends generally from an air-inlet side of the enclosure to an air-outlet side of the enclosure, and the fourth drive and a plurality of other drives in a fourth row of drives that extends generally from an air-inlet side of the enclosure to an air-outlet side of the enclosure.

5. The method of claim 1, further comprising: performing a third conversion of input DC power into a third regulated DC voltage source; performing a fourth conversion of input DC power into a fourth regulated DC voltage source; connecting the third regulated DC voltage source to the first drive and to the second drive but not to the third drive or the fourth drive; connecting the fourth regulated DC voltage source to the third drive and to the fourth drive but not to the first drive or the second drive.

6. The method of claim 1, further comprising: performing a third conversion of input DC power into a third regulated DC voltage source; performing a fourth conversion of input DC power into a fourth regulated DC voltage source; connecting the third regulated DC voltage source to the first drive and to the second drive but not to the third drive or the fourth drive; connecting the fourth regulated DC voltage source to the third drive and to the fourth drive but not to the first drive or the second drive; performing a first AC-to-DC conversion to generate a first source of input DC power and a second AC-to-DC conversion to generate a second source of input DC power both external to the enclosure; connecting the first source of input DC power from the first AC-to-DC conversion for the first regulated conversion of input DC power into the first regulated DC voltage source and for the fourth regulated conversion of input DC power into the fourth regulated DC voltage source; connecting DC power from the second AC-to-DC conversion for the second regulated conversion of input DC power into the second regulated DC voltage source and for the third regulated conversion of input DC power into the third regulated DC voltage source.

7. The method of claim 1, further comprising: performing a third conversion of input DC power into a third regulated DC voltage source; performing a fourth conversion of input DC power into a fourth regulated DC voltage source; connecting the third regulated DC voltage source to the first drive and to the second drive but not to the third drive or the fourth drive; connecting the fourth regulated DC voltage source to the third drive and to the fourth drive but not to the first drive or the second drive; performing a first AC-to-DC conversion to generate a first source of input DC power and a second AC-to-DC conversion to generate a second source of input DC power both external to the enclosure; connecting the first source of input DC power from the first AC-to-DC conversion for the first regulated conversion of input DC power into the first regulated DC voltage source and for the fourth regulated conversion of input DC power into the fourth regulated DC voltage source; connecting DC power from the second AC-to-DC conversion for the second regulated conversion of input DC power into the second regulated DC voltage source and for the third regulated conversion of input DC power into the third regulated DC voltage source; and sequencing the connecting of electrical power from the respective voltage sources to the respective disk drives such that the first disk drive is powered up before the second disk drive, and the third disk drive is powered up before the fourth disk drive.

8. An apparatus comprising: a single drive enclosure; a plurality of disk drives mounted in the enclosure including a first, a second, a third, and a fourth drive positioned in a pattern of disk drives that extends generally from an air-inlet side of the enclosure to an air-outlet side of the enclosure; a plurality of voltage regulators mounted in the enclosure including a first voltage regulator operatively coupled to receive input DC power and to output a first regulated DC voltage to the first drive and the second drive but not to the third drive or the fourth drive, and a second voltage regulator operatively coupled to receive input DC power and output a second regulated DC voltage to the third drive and the fourth drive but not to the first drive or the second drive; and a controller configured to write a first portion of a set of data to the first drive and a second portion of the set of data to the second drive and to write a mirrored copy of the first portion of the set of data to the third drive and a mirrored copy of the second portion of the set of data to the fourth drive such that if either the first or the second voltage regulator fails then both the first portion and the second portion of the set of data remain available from drives powered by the other non-failing voltage regulator.

9. The apparatus of claim 8, further comprising: a first AC-to-DC converter external to the enclosure and connected to the first voltage regulator to supply input DC power to the first voltage regulator; and a second AC-to-DC converter external to the enclosure and connected to the second voltage regulator to supply input DC power to the second voltage regulator.

10. The apparatus of claim 8, further comprising: a plurality of power routers including a first power router operatively coupled to selectively connect the first regulated DC voltage to the first drive, a second power router operatively coupled to selectively connect the second regulated DC voltage to the second drive, a third power router operatively coupled to selectively connect the third regulated DC voltage to the third drive, and a fourth power router operatively coupled to selectively connect the fourth regulated DC voltage to the fourth drive; and a sequencer the selectively connecting of power from the respective voltage sources to the respective disk drives such that the first disk drive is powered up before the second disk drive, and the third disk drive is powered up before the fourth disk drive.

11. The apparatus of claim 8, wherein the pattern of disk drives includes: the first drive and a plurality of other drives positioned in a first row of drives that extends generally from an air-inlet side of the enclosure to an air-outlet side of the enclosure, the second drive and a plurality of other drives positioned in a second row of drives that extends generally from an air-inlet side of the enclosure to an air-outlet side of the enclosure, the third drive and a plurality of other drives positioned in a third row of drives that extends generally from an air-inlet side of the enclosure to an air-outlet side of the enclosure, and the fourth drive and a plurality of other drives positioned in a fourth row of drives that extends generally from an air-inlet side of the enclosure to an air-outlet side of the enclosure.

12. The apparatus of claim 8, further comprising: a third voltage regulator operatively coupled to receive input DC power and to output a third regulated DC voltage to the first drive and the second drive but not to the third drive or the fourth drive; and a fourth voltage regulator operatively coupled to receive input DC power and output a fourth regulated DC voltage to the third drive and the fourth drive but not to the first drive or the second drive.

13. The apparatus of claim 8, further comprising: a third voltage regulator operatively coupled to receive input DC power and to output a third regulated DC voltage to the first drive and the second drive but not to the third drive or the fourth drive; a fourth voltage regulator operatively coupled to receive input DC power and output a fourth regulated DC voltage to the third drive and the fourth drive but not to the first drive or the second drive; a first AC-to-DC converter external to the enclosure and connected to the first and fourth voltage regulators to supply input DC power to the first and fourth voltage regulators; and a second AC-to-DC converter external to the enclosure and connected to the second and third voltage regulators to supply input DC power to the second and third voltage regulators.

14. The apparatus of claim 8, further comprising: a third voltage regulator operatively coupled to receive input DC power and to output a third regulated DC voltage to the first drive and the second drive but not to the third drive or the fourth drive; a fourth voltage regulator operatively coupled to receive input DC power and output a fourth regulated DC voltage to the third drive and the fourth drive but not to the first drive or the second drive; a first AC-to-DC converter external to the enclosure and connected to the first and fourth voltage regulators to supply input DC power to the first and fourth voltage regulators; a second AC-to-DC converter external to the enclosure and connected to the second and third voltage regulators to supply input DC power to the second and third voltage regulators; a plurality of power routers including a first power router operatively coupled to selectively connect the first regulated DC voltage to the first drive, a second power router operatively coupled to selectively connect the first regulated DC voltage to the second drive, a third power router operatively coupled to selectively connect the fourth regulated DC voltage to the third drive, and a fourth power router operatively coupled to selectively connect the fourth regulated DC voltage to the fourth drive; a fifth power router operatively coupled to selectively connect the third regulated DC voltage to the first drive, a sixth power router operatively coupled to selectively connect the third regulated DC voltage to the second drive, a seventh power router operatively coupled to selectively connect the second regulated DC voltage to the third drive, and a eighth power router operatively coupled to selectively connect the second regulated DC voltage to the fourth drive; and a sequencer that controls the selective connection of power from the respective voltage sources to the respective disk drives such that the first disk drive is powered up before the second disk drive, and the third disk drive is powered up before the fourth disk drive.

15. An apparatus comprising: a single drive enclosure holding a plurality of disk drives including a first, a second, a third, and a fourth drive positioned in a pattern of disk drives that extends generally from an air-inlet side of the enclosure to an air-outlet side of the enclosure; means for performing a first conversion of input DC power into a first regulated DC voltage source; means for performing a second conversion of input DC power into a second regulated DC voltage source; means for connecting the first regulated DC voltage source to the first drive and the second drive but not to the third drive or the fourth drive; means for connecting the second regulated DC voltage source to the third drive and the fourth drive but not to the first drive or the second drive; and means for writing a first portion of a set of data to the first drive and a second portion of the set of data to the second drive and a mirrored copy of the first portion of the set of data to the third drive and a mirrored copy of the second portion of the set of data to the fourth drive such that if either the first or the second conversion of input DC power to the respective first and second voltage source fails then both the first portion and the second portion of the set of data remain available from drives powered by the other non-failing voltage source.

16. The apparatus of claim 15, further comprising: means for performing a first AC-to-DC conversion to generate a first source of input DC power and a second AC-to-DC conversion to generate a second source of input DC power both external to the enclosure; means for connecting the first source of input DC power from the first AC-to-DC conversion for the first regulated conversion of input DC power into the first regulated DC voltage; and means for connecting DC power from the second AC-to-DC conversion for the second regulated conversion of input DC power into the second regulated DC voltage.

17. The apparatus of claim 15, further comprising: means for sequencing the connecting of power from the respective voltage sources to the respective disk drives such that the first disk drive is powered up before the second disk drive, and the third disk drive is powered up before the fourth disk drive.

18. The apparatus of claim 15, wherein the disk drives are positioned such that the pattern of disk drives includes: the first drive and a plurality of other drives in a first row of drives that extends generally from an air-inlet side of the enclosure to an air-outlet side of the enclosure, the second drive and a plurality of other drives in a second row of drives that extends generally from an air-inlet side of the enclosure to an air-outlet side of the enclosure, the third drive and a plurality of other drives in a third row of drives that extends generally from an air-inlet side of the enclosure to an air-outlet side of the enclosure, and the fourth drive and a plurality of other drives in a fourth row of drives that extends generally from an air-inlet side of the enclosure to an air-outlet side of the enclosure.

19. The apparatus of claim 15, further comprising: means for performing a third conversion of input DC power into a third regulated DC voltage source; means for performing a fourth conversion of input DC power into a fourth regulated DC voltage source; means for connecting the third regulated DC voltage source to the first drive and to the second drive but not to the third drive or the fourth drive; means for connecting the fourth regulated DC voltage source to the third drive and to the fourth drive but not to the first drive or the second drive.

20. The apparatus of claim 15, further comprising: means for performing a third conversion of input DC power into a third regulated DC voltage source; means for performing a fourth conversion of input DC power into a fourth regulated DC voltage source; means for connecting the third regulated DC voltage source to the first drive and to the second drive but not to the third drive or the fourth drive; means for connecting the fourth regulated DC voltage source to the third drive and to the fourth drive but not to the first drive or the second drive; means for performing a first AC-to-DC conversion to generate a first source of input DC power and a second AC-to-DC conversion to generate a second source of input DC power both external to the enclosure; means for connecting the first source of input DC power from the first AC-to-DC conversion for the first regulated conversion of input DC power into the first regulated DC voltage source and for the fourth regulated conversion of input DC power into the fourth regulated DC voltage source; means for connecting DC power from the second AC-to-DC conversion for the second regulated conversion of input DC power into the second regulated DC voltage source and for the third regulated conversion of input DC power into the third regulated DC voltage source.

21. The apparatus of claim 15, further comprising: means for performing a third conversion of input DC power into a third regulated DC voltage source; means for performing a fourth conversion of input DC power into a fourth regulated DC voltage source; means for connecting the third regulated DC voltage source to the first drive and to the second drive but not to the third drive or the fourth drive; means for connecting the fourth regulated DC voltage source to the third drive and to the fourth drive but not to the first drive or the second drive; means for performing a first AC-to-DC conversion to generate a first source of input DC power and a second AC-to-DC conversion to generate a second source of input DC power both external to the enclosure; means for connecting the first source of input DC power from the first AC-to-DC conversion for the first regulated conversion of input DC power into the first regulated DC voltage source and for the fourth regulated conversion of input DC power into the fourth regulated DC voltage source; means for connecting DC power from the second AC-to-DC conversion for the second regulated conversion of input DC power into the second regulated DC voltage source and for the third regulated conversion of input DC power into the third regulated DC voltage source; and means for sequencing the connecting of electrical power from the respective voltage sources to the respective disk drives such that the first disk drive is powered up before the second disk drive, and the third disk drive is powered up before the fourth disk drive.

Description:

CROSS-REFERENCES TO RELATED INVENTIONS

This application is additionally related to:

U.S. patent application Ser. No. 11/537,600 filed on Sep. 29, 2006 and entitled “DISK-DRIVE ENCLOSURE HAVING FRONT-BACK ROWS OF SUBSTANTIALLY PARALLEL DRIVES AND METHOD”;
U.S. patent application Ser. No. 11/537,605 filed on Sep. 29, 2006 and entitled “DISK-DRIVE ENCLOSURE HAVING ROWS OF ALTERNATELY FACING PARALLEL DRIVES TO REDUCE VIBRATION AND METHOD”;
U.S. patent application Ser. No. 11/537,606 filed on Sep. 29, 2006 and entitled “DISK-DRIVE ENCLOSURE HAVING LATERALLY OFFSET PARALLEL DRIVES TO REDUCE VIBRATION AND METHOD”;
U.S. patent application Ser. No. 11/537,608 filed on Sep. 30, 2006 and entitled “DISK-DRIVE ENCLOSURE HAVING NON-PARALLEL DRIVES TO REDUCE VIBRATION AND METHOD”;
U.S. patent application Ser. No. 11/537,614 filed on Sep. 30, 2006 and entitled “DISK-DRIVE ENCLOSURE HAVING PAIR-WISE COUNTER-ROTATING DRIVES TO REDUCE VIBRATION AND METHOD”;
U.S. patent application Ser. No. 11/537,598 filed on Sep. 29, 2006 and entitled “DISK-DRIVE ENCLOSURE HAVING DRIVES IN A HERRINGBONE PATTERN TO IMPROVE AIRFLOW AND METHOD”;
U.S. patent application Ser. No. 11/537,607 filed on Sep. 30, 2006 and entitled “DISK-DRIVE SUPPORTING MASSIVELY PARALLEL VIDEO STREAMS AND METHOD”;
U.S. patent application Ser. No. 11/537,610 filed on Sep. 30, 2006 and entitled “DISK-DRIVE SYSTEMS WITH VARYING NUMBER OF SPARES FOR DIFFERENT EXPECTED LIFETIMES AND METHOD”;and
U.S. patent application Ser. No. 11/537,613 filed on Sep. 30, 2006 and entitled “PORUOS LIGHT-EMITTING DISPLAY WITH AIR FLOW THROUGH DISPLAY, ITS USE IN A DISK-DRIVE SYSTEM AND METHOD”; which are all hereby incorporated by reference in their entirety.

FIELD OF THE INVENTION

This invention relates to hard-disk-drive data-storage systems and methods, and more particularly, to enclosures that hold a large number of hard-disk-drives and provide a large number of serial data interfaces operating in parallel, resulting in, among other things, improved performance, reliability, manufacturing costs, and/or operational costs.

BACKGROUND OF THE INVENTION

Massive amounts of data storage are required for many emerging and existing applications. For example, video-on-demand applications can provide access to hundreds or thousands of movies for hundreds or thousands of users simultaneously, requiring vast amounts of digital storage, fast access, 24 hours-per-day and 7 days per week (24/7) availability and uptime, and huge bandwidth. Modern supercomputers also require these features, as well as requiring even faster access, extraordinary data integrity, error checking and error correction.

Semiconductor memories provide very fast access, reasonable densities, and moderate costs. However, most common semiconductor memories are volatile (they lose their data when not powered or not refreshed on a timely basis), they develop soft errors (errors that can be corrected by re-writing the affected location) due to various causes including alpha radiation, and they can be cost prohibitive. Additionally, the heat and power requirements can be problematic, if they are used to store vast amounts of information for long time periods.

Hard-disk drives (HDDs, also called just “disk drive” or “drive”)) provide cost-effective non-volatile data storage on rotating media. Data are written and read by magnetic transducer heads that are moved to one of thousands of tracks to locate requested data. There are time penalties incurred to move the head to the requested track, to rotate the disk to the requested location on that track, and to serially read or write the data from or to the track location. The moving parts of a disk drive are prone to wear and failure over time. For applications requiring high reliability (error-free data) and availability (24/7 uptime), data can be stored in a redundant manner (e.g., redundant arrays of inexpensive disks, or RAID), and several different RAID schemes are known to the art, frequently making compromises between performance, cost, and data recoverability. Another requirement for many applications is serviceability—the ease of repairing a faulty system in the field (i.e., at a customer's location of the equipment).

Data storage servers (enclosures having one or more disk drives as well as a data processor to receive data access requests and control the storing and fetching of data to and from the disk drives) and storage vaults (enclosures having one or more disk drives but essentially no processor, and using a data processor housed in a separate enclosure to receive data access requests and control the storing and fetching of data to and from the disk drives) can be implemented in free-standing units (typically an upright unit placed on the floor or on a desk) or as rack-mount units (typically horizontally-oriented units bolted to a standardized nineteen-inch (48.26 cm) rack).

Typical conventional rack-mount disk-drive enclosures arrange a plurality (3 to 14) HDDs in removable carriers that are accessible from the “front” of the unit (the side typically facing a user area), and usually are arranged so that data and power cables are accessible from the “back” of the unit. The disk drives can thus be replaced fairly easily if one were to fail. RAID solutions can be utilized to use redundant data artifacts to compute the data that was on the failed disk drive. This data is sent to a requestor or used to recreate the data on a new (spare) disk drive once one is inserted to replace the failed unit. Since racks of rack-mount units are often installed in rows, there is typically no access provided from the sides of a rack-mount unit, and since the rack-mount units are stacked one on top of another in each rack there is typically no access provided from the top or bottom of a rack-mount unit.

High-density packaging of HDDs in an enclosure exacerbates drive-to-drive vibration interaction problems. With several HDDs, packaged closely together in single enclosure, potentially many doing simultaneous head-seeks, the vibration interaction problem is greatly increased. Previous systems and methods to package HDDs and reduce drive-to-drive vibration interaction involved mechanical stiffening of the enclosure and/or lower density packaging options.

Numerous computer applications utilize multiple disk drives for data storage and acquisition. These multiple disk drives are often located in separated locations. For example, disk drives may be arranged in rack systems that consume large amounts of space and require multiple cabinets to house the rack systems. Furthermore, positioning multiple disk drives in separate locations adds to the complexity of data acquisition from the disk drives because a more complex interface with the multiple disk drives is required. In addition, longer cabling is required to reach the separately located disk drives. Accordingly, what is needed is an apparatus that positions multiple disk drives in a manner that simplifies data acquisition from the disk drives and reduces the space needed to house the multiple disk drives.

SUMMARY OF THE INVENTION

In some embodiments, the present invention generally involves housing a large number of disk drives in an enclosure. In other embodiments, the invention is based on positioning disk drives such that forces occurring during seek and write functions within a first disk drive are counteracted by analogous forces occurring in one or more other drives that are positionally paired with the first disk drive in some embodiments. An example of such a force includes rotation and counter-rotation of disks that is caused by movement of an actuator arm within the disk drive that occurs during a seek or write function of the disk. Other examples of such forces include vibrational forces, rotational, counter-rotational forces, and the like that are due to the movement of a disk within a disk drive. These forces can be caused by numerous actions within a disk drive. Arranging the disk drives according to the invention helps to reduce detrimental results caused by such forces that can increase the incidence of read and write errors. Accordingly, the invention can be used to position multiple disk drives so that the disk drives have a reduced read and write error rate.

In some embodiments, the invention provides an apparatus that includes a substrate, and a plurality of disk drives each coupled electrically and mechanically to the substrate, the plurality of disk drives including at least a first and a second disk drive, wherein the first disk drive is positioned relative to the second disk drive so that a rotational force produced by the first disk drive is at least partially counteracted by a rotational force produced by the second disk drive.

In other embodiments, the invention provides a method that includes mounting a plurality of drives in an enclosure, the enclosure including a connector substrate, the plurality of drives including at least a first disk drive and a second disk drive that are each electrically and mechanically coupled to the enclosure; and mechanically coupling the first drive and the second drive such that rotational force produced by the first disk drive is at least partially counteracted by rotational force produced by the second disk drive.

In some embodiments, the invention provides an apparatus that includes an enclosure that includes a substrate, a means in the enclosure for mounting a plurality of disk drives to the enclosure, and a means for coupling a plurality of disk drives electrically and mechanically to the substrate, the plurality of disk drives including at least a first and a second disk drive, wherein the first disk drive is positioned relative to the second disk drive so that a rotational force produced by the first disk drive is at least partially counteracted by a rotational force produced by the second disk drive.

In some embodiments, the invention provides an apparatus that includes a substrate, and a plurality of disk drives each coupled electrically and mechanically to the substrate, the plurality of disk drives including at least a first disk drive and a second disk drive, wherein the first and second disk drive each have a first major face surrounded by a first, second, third and fourth edge and having a first, second, third and fourth corner, wherein the first disk drive and the second disk drive are positioned such that a rotational force produced by the first disk drive is conveyed primarily as a translational force to the second disk drive.

In some embodiments, the invention provides a method that includes mounting a plurality of drives in an enclosure, the plurality of drives including at least a first disk drive and a second disk drive that are each electrically and mechanically coupled to the enclosure, and mechanically coupling the first disk drive and the second disk drive such that rotational force produced by the first disk drive is at least partially transmitted as translational force to the second disk drive.

In some embodiments, the invention provides an apparatus that includes a substrate; and a means for mounting a plurality of disk drives to the substrate; and a means for coupling a plurality of disk drives electrically and mechanically to the substrate, the plurality of disk drives including at least a first disk drive and a second disk drive, wherein the first and second disk drive each have a first major face surrounded by a first, second, third and fourth edge and having a first, second, third and fourth corner, wherein the first disk drive and the second disk drive are positioned such that a rotational force produced by the first disk drive is conveyed primarily as a translational force to the second disk drive.

In some embodiments, the invention provides an apparatus that includes a substrate, and a plurality of disk-drive connectors each coupled electrically and mechanically to the substrate, the plurality of disk-drive connectors including at least a first and a second disk-drive connector, wherein the first disk-drive connector is positioned relative to the second disk-drive connector so that a rotational force produced by a first disk drive that is connected to the first disk-drive connector is at least partially counteracted by a rotational force produced by a second disk drive that is connected to the second disk-drive connector.

In some embodiments, the invention provides an apparatus that includes a substrate, and a plurality of disk-drive connectors each coupled electrically and mechanically to the substrate, the plurality of disk-drive connectors including at least a first disk-drive connector and a second disk-drive connector, wherein the first disk-drive connector and the second disk-drive connector are positioned such that a rotational force produced by a first disk drive connected to the first disk-drive connector is conveyed primarily as a translational force to a second disk drive connected to the second disk-drive connector.

In some embodiments, the invention provides a method that includes mounting a plurality of disk-drive connectors in an enclosure, the enclosure including a connector substrate, the plurality of disk-drive connectors including at least a first disk-drive connector and a second disk-drive connector that are each electrically and mechanically coupled to the enclosure, and mechanically coupling the first disk-drive connector and the second disk-drive connector such that rotational force produced by a first disk drive that is connected to the first disk-drive connector is at least partially counteracted by rotational force produced by a second disk drive that is connected to the second disk-drive connector.

In some embodiments, the invention provides a method that includes mounting a plurality of disk-drive connectors in an enclosure, the plurality of disk-drive connectors including at least a first disk-drive connector and a second disk-drive connector that are each electrically and mechanically coupled to the enclosure, and mechanically coupling the first disk-drive connector and the second disk-drive connector such that rotational force produced by a first disk drive that is connected to the first disk-drive connector is at least partially transmitted as translational force to a second disk drive that is connected to the second disk-drive connector.

In some embodiments, the invention provides a method that includes mounting a plurality of disk drives in an enclosure, the enclosure including a connector substrate, the plurality of disk drives including at least a first disk drive and a second disk drive; vibrationally coupling the first disk drive to the second disk drive, and sending a first seek operation to the first disk drive and a second seek operation to the second disk drive, wherein a timing of the first seek operation relative to the second seek operation is adjusted to minimize adverse vibrational interaction between the first disk drive and the second disk drive.

In some embodiments, the invention provides an apparatus that includes a data structure having a plurality of entries, each entry containing vibration-interaction information relative to a read operation occurring on a first disk drive of a pair of disk drives and a seek operation being performed on a second disk drive of the pair.

In some embodiments, the invention provides an apparatus that includes a memory, the memory holding vibration-interaction information, an information processing unit operatively coupled to the memory to receive the vibration-interaction information and adjusting a timing of seek operations to a plurality of disk drives based on the information.

In some embodiments, the invention provides a method that includes mounting a plurality of disk drives in shock mounts in an enclosure and “detenting” the plurality of disk drives against vibration using a disengagable detent device.

In some embodiments, the invention provides an apparatus that includes an enclosure, a substrate held within the enclosure, a plurality of disk-drive connectors each coupled mechanically to the substrate, the plurality of disk-drive connectors including at least a first and a second disk-drive connector, and an over-shock detector operatively coupled to the enclosure and adapted to detect and store information regarding one or more over-shock events.

In some embodiments, the invention provides a method that includes analyzing vibration-interaction between a plurality of disk drives held in an enclosure and storing information that is based on the analysis into a data structure.

In some embodiments, the invention provides a method that includes mounting a plurality of disk drives to disk-drive connectors within an enclosure, adhering a resilient sheet across the plurality of disk drives, and attaching a cover to the resilient sheet.

In some embodiments, the invention provides an apparatus that includes a plurality of disk drives mounted to disk-drive connectors within an enclosure, a resilient sheet (such as a visco-elastic membrane, for example) across the plurality of disk drives, and a cover.

BRIEF DESCRIPTION OF THE DRAWINGS

Various features and attendant advantages of the present invention will become fully appreciated as the invention becomes better understood upon reading the following description and when considered in conjunction with the accompanying drawings, in which like reference characters designate the same or similar parts throughout the several views.

FIG. 1 is a perspective drawing of disk drive 120 mounted in a perpendicular-to-the-major-face orientation (e.g., vertical, if the major face is horizontal) in a disk-drive system 100.

FIG. 2 is a perspective drawing of a storage system 200 with the disk drives placed in a new physical-layout pattern 250 that enables the disk drives themselves to serve as the “fins” of a large heat sink.

FIG. 3A is a block diagram of a power supply 300, as used in some embodiments.

FIG. 3B is a block diagram of a power supply 300′, as used in some embodiments.

FIG. 3C is a block diagram of a power supply 300″, as used in some embodiments.

FIG. 4A is a perspective drawing of disk drives 120 and 120′ mounted in a vertical orientation in a disk-drive system 100.

FIG. 4B is a perspective drawing of a pair of disk drives in a T orientation.

FIG. 4C is a perspective drawing of a pair of disk drives in a Y orientation.

FIG. 4D is a perspective drawing of a pair of disk drives in a counter-rotating parallel orientation with their axes of rotation aligned.

FIG. 4E is a perspective drawing of a pair of disk drives in a counter-rotating parallel orientation with their edges aligned.

FIG. 4F is a perspective drawing of a pair of disk drives in a counter-rotating parallel orientation each with its axis of rotation aligned with an edge of the other disk drive.

FIG. 4G is a plan-view schematic of a herringbone configuration 400′ with counter-rotating pairs of disk drives.

FIG. 5 is a plan-view schematic of a herringbone configuration 500 with counter-rotating pairs of disk drives.

FIG. 6A is a plan-view schematic of another herringbone configuration 600 with counter-rotating pairs of disk drives.

FIG. 6B is a plan-view schematic of another herringbone configuration 601 with counter-rotating pairs of disk drives.

FIG. 7A shows a plan view of yet another herringbone configuration 700 of disk drives.

FIG. 7B shows a perspective view of system 700.

FIG. 8A is a perspective drawing of prior-art “high-density” hard-disk-drive (HDD) enclosure systems 81 and 82 as might be mounted in a rack 80.

FIG. 8B is a perspective drawing of a high-density HDD enclosure system 810 according to the present invention.

FIG. 8C is a perspective drawing of a high-density HDD enclosure system 811 using a herringbone configuration according to the present invention.

FIG. 8D is a perspective view that illustrates a perforated support grid for a plurality of disk drives with ESD-(electro-static discharge prevention)-coated visco-elastomeric material.

FIG. 8E is a top view that illustrates nesting support grid for a plurality of disk drives with ESD-(electro-static discharge prevention)-coated visco-elastomeric material.

FIG. 8F is a perspective view that illustrates system 804 having a molded-in connector 819 support for a plurality of drives mounted in a vertical orientation.

FIG. 8G is a top view of system 804 of FIG. 8F.

FIG. 8H is top view that illustrates the distribution of temperature sensors around the inlet manifold 1112, outlet manifold 1114 and between-drive spaces 95.

FIG. 8I is a front view that illustrates the status-display grid 816.

FIG. 8J is a perspective view that illustrates a cover-latching mechanism that seats the drives into their connectors.

FIG. 9A is a perspective view that illustrates a porous display having LEDs mounted on a screen that has much space for air flow through the displays.

FIG. 9B is a perspective view that illustrates an LCD display mounted on the inlet air dams allowing much space for air flow around the displays.

FIG. 9C is a front-elevation view that illustrates an LCD display mounted on the inlet air dams allowing much space for air flow around the displays.

FIG. 10 is a blown-up perspective view of a system 1000 of some embodiments having one or more disk-drive systems 1001 operatively coupled to one or more central processing units (CPU) 1002 and/or one or more video-streaming units 1003 or some combination thereof.

FIG. 11 is a plan-view block diagram of a data-storage system 1100 of some embodiments of the invention that provides a high density enclosure having one or more rows of disk drives.

FIG. 12 is a plan-view block diagram of a data-storage system 1200 of some embodiments of the invention that uses tapered inlet and outlet air chambers.

FIG. 13 is a plan-view block diagram of a data-storage system 1300 of some embodiments of the invention that uses curving tapered inlet and outlet air chambers.

FIG. 14 is a plan-view block diagram of a data-storage system 1400 of some embodiments of the invention that uses curving tapered inlet and outlet air chambers, and laterally offset paired drives.

FIG. 15 is a plan-view block diagram of a connector circuit card pair 1500 used in some embodiments of the invention.

FIG. 16A is a plan-view block diagram of a data-storage system 1600 of some embodiments of the invention that provides a high density enclosure having four rows of disk drives.

FIG. 16B is a functional block diagram of a circuit 1608 used in some embodiments of system 1600.

FIG. 16C is a functional block diagram of a circuit 1609 used in some embodiments of system 1600.

FIG. 17 is a plan-view block diagram of a data-storage system 1700 of some embodiments of the invention that provides a high density enclosure having one or more rows of disk drives accommodating a variable number of disk drives in each row.

FIG. 18 is a perspective-view block diagram of a data-storage system 1800 of some embodiments of the invention that provides one or more rows of disk drives in an upper portion of the enclosure and one or more power supplies in an adjacent lower portion of the enclosure.

FIG. 19 is an elevation view of a data-storage system 1900 of some embodiments of the invention that provides a high-density enclosure having one or more rows of disk drives.

FIG. 20A is an elevation view of a data-storage system 2000 of some embodiments of the invention that provides a high-density enclosure having one or more rows of disk drives arranged in coupled pairs of counter-rotating disk drives.

FIG. 20B is an elevation view of a data-storage system 2001 of some embodiments of the invention that provides a high-density enclosure having one or more rows of disk drives with an adjustable-height mid-drive vibration damper 2075.

FIG. 20C is an elevation view of a data-storage system 2002 of some embodiments of the invention that provides a high-density enclosure having one or more rows of disk drives with a cast-in-place vibration-damper boot 2076.

FIG. 20D is an elevation view of a data-storage system 2003 of some embodiments of the invention that provides a high-density enclosure having one or more rows of disk drives with a cast-in-place mid-drive vibration damper 2077.

FIG. 21 is a front elevation view of a data-storage system 2100 of some embodiments of the invention that provides a high-density enclosure having one or more rows of disk drives with vertical beam stiffener 2110 and optional vibration damper 2122.

DETAILED DESCRIPTION

In the following detailed description of preferred embodiments, reference is made to the accompanying drawings that form a part hereof, and in which are shown, by way of illustration, specific embodiments in which the invention may be practiced. It is to be understood that other embodiments may be utilized and structural changes may be made without departing from the scope of the present invention. The references to relative terms such as top, bottom, upper, lower, vertical, horizontal, etc., refer to an example orientation such as used in the Figures, and not necessarily an orientation used during fabrication or use.

Systems and methods to densely package disk drives in an enclosure, while at the same time reducing negative effects on the disk drives that are due to drive-to-drive interactions, can improve performance, density, reliability, and also reduce manufacturing costs and operational costs.

Individual disk drives include one or more head-disk assemblies (HDAs) and the electronics for control and data transfer to and from the disks. The HDA includes one or more disks and one or more actuator on which a head is attached. An actuator to which a head is attached is positioned within the disk drive such that the actuator can be rotated about an axis to selectively position the attached head to a select location on an adjoining disk. Accordingly, data can be retrieved from, or written to, a specific location on a disk by movement of the actuator to position the attached head at the specific location on the disk.

System Environment

The present invention provides improved systems and methods to densely package the hard-disk drives in an enclosure, while at the same time reducing drive-to-drive vibration interaction. These can improve performance, density, reliability, and also reduce manufacturing and operational costs. Each hard-disk drive (HDD, also called “disk drive” or “drive”) includes one or more HDAs and the electronics for control and data transfer to and from the disks.

Hard-disk-drives are sensitive to vibration. The performance and reliability of a HDD are decreased with vibration. When multiple HDDs are operating within an enclosure, rotational-acceleration vibration generated from the head-seek operation on one HDD can adversely affect the read/write operations (and possibly head-seek operations as well) on other HDDs. (Note that non-acceleration vibration such as due to disk-spindle wobble, room noise, or fan vibration is generally less problematic than acceleration vibration due to actuator seek operations.) The drive-to-drive rotational-acceleration vibration interaction can cause the heads in an HDD to move off track, and thus cause read-data errors and write-data errors. Such errors may result in additional revolutions to re-locate the data, excessive retries, lost data, longer head-seek times, slow data access, increase power consumption and heat production. Reducing the vibration transferred between HDDs can improve HDD performance, density, reliability, manufacturing costs, and/or operational costs.

FIG. 1 is a perspective drawing of disk drive 120 mounted in a perpendicular-to-the-major-face orientation (e.g., vertical, if the major face is horizontal) in a disk-drive system 100. In some embodiments, a plurality of other drives (up to one-hundred-fifty, one-hundred-ninety-two, two-hundred or two-thousand drives or more) are each plugged into their respective sockets (or to other suitable connectors) (e.g., connector 123) that are coupled to connector circuit 129 (e.g., in some embodiments, a plurality of insulated conductors carrying power and signals to and from drive 120) on connector circuit board or substrate 150. Disk drive 120 includes one or more disks 115 that rotate around their axis 117, an actuator 112 that rotates back and forth around its axis 111 to move its head 114 onto a given track 113 on disk 115. The data is written serially on each track 113 (e.g., as magnetic domains in the case of magnetic recording disks, or as optical artifacts in the case of optical disks, or as atomic-force artifacts or other suitable information), so the head 114 must be moved to and kept on track 113 in order to read the data. Any movement of drive 120 that causes the drive 120 to have a rotational force 187 around its Z_R120center-of-mass axis, or a transitional rotation vibration force, can cause head 114 to be moved off track 113.

Data is organized on the disk drive 120 in serial fashion. This means that the data is stored on individual tracks (e.g., track 113) on the disk 115, which can be exemplified as concentric rings. A head that is positioned at a constant radius from the center of rotation of the disk is able to read data from a specific track on the disk as the disk turns. This allows data to be stored and retrieved from specific tracks on the disk by positioning the head above the specific track. However, if the position of the head is disrupted (i.e., moved off track), the head is no longer able to read the data from the desired specific track and must be repositioned. Accordingly, events that cause the position of the head to change in an undesired manner disallow proper reading of data from a disk and disallow proper writing of data to the disk. Examples of such events include shock to the disk drive, vibrational forces, torques, and the like.

The time required to find and transfer data on a disk is referred to as the access time. Access time can be divided into seek time, rotational latency, and data transfer time. Seek time refers to the time required to position an actuator on a track that contains the desired data. Rotational latency refers to the time required for the disk to spin such that the desired data on the requested track is under the head 114 of the properly positioned actuator 112. Transfer time refers to the time required to transfer the data to or from the head 114 on the actuator 115 to a location on a track 113 where the data is stored or retrieved (put to use). The rate of data transfer can be altered by placing different portions of the data on different disk drives (this is called striping, explained further below). For example, data can be split into blocks that are stored on two or more disk drives. Different blocks of data can then be read from the multiple disk drives in an overlapped or parallel manner and used as needed without having to wait for a single disk drive to free up. This process allows overall data to be transferred more rapidly than if the data are stored on a single disk drive. The rate of data acquisition can also be altered by placing multiple copies of data onto a disk. For example, five copies of the same data block can be stored on a single track or closely adjacent tracks of a disk to reduce rotational latency as the disk would only have to turn at most one-fifth of a revolution for one of the copies of the data to be accessed (one tenth of a revolution on average), as compared to accessing data on average in one-half revolution for data that was stored on the disk as a single copy. (Since the location on the track where the head starts is random with respect to the location of the data, some of the time the head will reach the track exactly at a point in time that it can immediately access the data (no revolution time), and other times it will take a full revolution until the data is in a position to be accessed; thus, on average the rotational latency is generally a half revolution is a single copy of the data is used, and ½N revolutions if N copies of the data are stored.) Additionally, storing multiple copies of data on a single track can decrease the time required for data acquisition in the event of a tracking or other recoverable error, since the rotational latency would be reduced following repositioning of the head following the error.

When data is retrieved or written to a disk, a seek operation is used that rotates the actuator about its axis and positions the attached head at the track on the disk where the data is to be written or read. The rotation of the actuator arm produces a rotational force, wherein the disk drive experiences a rotational force in the opposite direction as the actuator motion. This rotational force can move the disk drive and thus move the neighboring enclosure and cause a neighboring drive to move. This can cause the track position of the actuator in that neighboring disk drive to change and if that disk drive is reading or writing data at the time, it will thereby cause a read or write error to occur in the neighboring drive.

In conventional disk-drive arrays, the enclosure and the HDA cases are quite heavy in relation to the mass of the actuator. Accordingly, the disk drives of the disk-drive arrays are less affected by rotational forces that are transferred from one disk drive doing a seek operation to a neighboring disk drive doing a read or write operation. As the mass of the HDA is reduced, the proportional mass of the actuator increases, and the relative rotational force due to the actuator is relatively larger. In addition, smaller drives allow the enclosure's metal case (which is used to fabricate the disk-drive-array enclosure) to be made thinner and less rigid. The resulting lighter weight can produce less damage to the unit if it is dropped. However, the thinner metal can also allow a greater amount of rotational or translational force to be transmitted between drives. Generally, moderate translational force is not a problem, nor is rotational force that does not move the read-write head (e.g., rotational acceleration around an axis perpendicular to the actuator axis). With increasingly smaller drives and thinner cases, the rotational force from a seek operation in one drive has a larger deleterious affect (i.e., primarily a rotational force that moves a head off track) that is transmitted to nearby disk drives and that results in the problems described.

Accordingly, these negative effects of rotational and translational force on disk drives are exacerbated by two major trends in the disk drive and disk array industries. The first of these is the trend toward smaller and lighter HDA mechanisms. As HDA mechanisms become smaller (as a function of disk diameter), the mass of platters decreases roughly as a function of the square of the platter radius. The mass of disk drive motors also tends to decrease exponentially as a function of disk diameter. However the mass of the head actuator tends to decrease only linearly, as a function of the length of the actuator. The result is that as HDA mechanisms become smaller, the mass of the actuator becomes a proportionately larger part of total HDA mass. The non-actuator portion of total HDA mass acts (beneficially) as an inertial mass (i.e., a damper of higher frequency vibrations since the heavier mass has a lower characteristic frequency) that attenuates rotational force, so the loss of non-actuator mass in proportion to total HDA mass represents a growing problem in disk arrays.

The second of these is the trend in disk arrays toward larger numbers of disk drives per unit of disk enclosure volume. Conventionally, these drives are lined up along the narrow front and/or back surface of the enclosure, where the right-angle corners constrain rotation and/or vibration. As disk drives are packaged more densely, they must be mounted interior to the enclosure on the membranes formed by the lower and/or upper covers, and the effect of inter-drive mechanical coupling and rotational and translational forces to nearby disk drives is exacerbated. With high-density enclosures and random disk accesses, the possibility of several HDDs generating additive rotational and/or translational forces is increased. In addition, the problem is greatly magnified for HDDs attempting to hold sector tracking while doing reads or writes.

FIG. 2 is a perspective view that illustrates a storage system 200, according to some embodiments of the invention, with the disk drives placed in a new physical-layout pattern 250 that enables the disk drives 120, 120′, and disk-drive pairs 205, 206, 207, 208, 209, (each having two disk drives 120) and the like, to individually and collectively serve as the “fins” of a large heat sink through which air is drawn or pushed in order to remove heat generated by the disk drives and the driving circuitry connected to use the disk drives. The arrangement of the disk drives further creates a plurality of tuned spaces such as inlet manifold 1112, outlet manifold 1114 and between-drive spaces 95 that control air flow from fans 240 to a high degree of precision in order to increase cooling efficiency. In some embodiments, the staggered herringbone orientation of HDDs with graduated spacing between disk drives is to optimize cooling by forcing airflow between the disk drives and taking into account the increasing temperature of the air as it moves through the disk drives. Since heat transfer is proportional to the temperature difference between the air and the drives, and to the amount of air, more air is used where the air temperature is higher and the temperature difference is less. In some embodiments, system 200 is connected to one or more processors 89, each coupled to communicated data to a plurality of disk-drive enclosure systems 201, 202, and/or 203 and the like, each having a large plurality of disk drives 120. In some embodiments, two or more power supplies 231, 232 provide redundant power for the disk drives 120. In some embodiments, the fans 240 are locates at a far end of the airflow through the enclosure so they pull air through the disk drives and push the heated air out of the cabinet in order that the heat from the fans is inserted into the air stream after it has cooled the other components. In some embodiments, the fans 240 are accessible and possibly replaceable by the user or service persons at an exterior surface of the enclosure, but enough redundancy is provided for the disk drives and power so that the system can continue to operate with substantially full functionality even if multiple individual components fail. Thus, the disk drives can be held in place in the enclosure using visco-elastic adhesive along one or a few edges, reducing weight and virtually eliminating the need for service calls. Further, small DC-to-DC regulated power supplies can be permanently mounted (e.g., soldered, in order to reduce connector-caused failures) in place, since multiple ones of the power supplies can fail and yet the system continues to function fully using the remaining good power supplies.

Power-Supply Description

FIG. 3A shows a disk drive system 201 having a power supply 300, as used in some embodiments of the invention. Power supply 300 includes a power crossover and power router configuration that meets the needs of a dense box of disk drives (DBOD). Power supply 231 includes two DC-to-DC power supplies 231A and 231B.

In some embodiments, each of these uses a AM80A-048L-050F40 model power supply available from Astec company. In some embodiments, the input to such a power supply includes dual 48-Volt DC supply lines with optional remote-control telecommunications to control the power. In some embodiments, the power modules can take DC input power from 36- to 72-Volt DC. One or more of the following features apply to some embodiments of the invention. The PRIMARY and MIRROR notation refers to drives that provide the primary data storage (the primary copy of data) and the mirrored data storage (the other copy or copies of the data). In some embodiments, there is no difference between primary and mirror copies of data, in that all write operations will write to all copies of the data, and read operations will only access one of the copies, wherein the selection of which copy is to be read is made on a rotation or alternating basis, or on a basis of which disk drive is not busy with another operation at the time when the read operation is started. For example, if the data are mirrored three ways, three disk drives will each have a copy of the same data, and when writing, the write data will be sent to all three disk drives, but when reading, a first read operation is sent to only the first disk drive, a second read operation is sent to only the second disk drive, and a third read operation is sent to only the third disk drive. When a fourth read operation arrives, it would generally be sent to the first disk drive, but if that disk drive is still busy with the first read operation, the fourth read operation could be sent to the second or third disk drive if either of those were finished with their earlier operations. By spreading the read operations among all the drives, it is more likely that a drive with the requested data for a particular read request will be available (that the data is on a drive that is not already busy with another prior operation).

In some embodiments, “Power Module Redundancy” is provided on the input, (i.e., each disk drive is configured to receive power from each of two or more DC-to-DC power supplies) wherein if any DC-to-DC power supply fails, it can be automatically disconnected and the remaining DC-to-DC power supply or supplies is able to handle the load. Like aircraft engines that have two spark plugs per cylinder, four cylinders, and “crossed over” ignitions for redundancy (e.g., two-way), some embodiments of the invention take a similar approach. In some embodiments, the sources 48V A and 48V B also cross the primary and mirrored boundaries. Dual redundant input (of the 48-volt DC sources) and the crossover configuration provide capability to power both sides in the event of a single 48V input loss. Each input can power both sides. In some embodiments, the power modules are made by Astec and provide less than 100-mV ripple (which is, in some embodiments, a requirement for the disk drives and some other power supplies cannot meet this), are parallelable, controllable, provide monitor sensors (e.g., voltage, temperature and current), provide high reliability that is more than one million hours MTBF (mean time between failures), regulatory approvals, and provide four voltage-range options: 18-36 VDC, 36-72 VDC, 90-200 VDC, and 180-400 VDC. This allows some embodiments to obtain power simply from AC, for example using a simple rectifier on the front end. In some embodiments, these power supplies provide an efficiency of 84 percent typical for 5 volts output, and ripple is 50 mV typical, and maximum 100 mV. In some embodiments, the entire box or enclosure of a plurality of drives is made to be “Hot Box” swappable (i.e., where an entire subsystem box is swapped out while the system is running), with just a little more switching to selectably disconnect power supplies 231 and 232 from their power sources.

In some embodiments, the next section or stage is the “power router.” This is a plurality of high-current, redundant relays (having a relatively low voltage drop at high current as compared to solid-state relays that have higher voltage drops) that can interconnect with each other, or switch power around, providing routing (if one should fail). When no power supply has failed, the switches connect a plurality of power supplies to each section of disk drives, thus reducing the amount of power that must be supplied by each power supply (e.g., in normal mode, each power supply provides half the power needed, and once a power supply fails, the other power supply provides all the power for its disk drives).

The last stage includes the disk drives. In some embodiments, each disk drive uses 5 volts DC, 5.5 Watts maximum (less than about one amp during power up). Lines drawn that “Link” the disk drives indicate which drives are mirrored, in some embodiments. This provides a data link between various copies of the mirrored data across different power sources 48-volt source A and 48-volt source B. In some embodiments, battery-backed uninterruptible power supplies (UPS) are provided for these sources. In some embodiments, Astec AM80A modules produce 240 Watts at 5-Volts DC, or 40 Amps at 5-Volts DC, for a 48-VDC input. In some embodiments, a version is used that is pin for pin compatible but more expensive, BM80A, 300W, 60A, if a design needs more power.

Some embodiments include four rows of forty-eight disk drives for a total of one-hundred-ninety-two drives. Rows are powered up one row at a time, sequentially over a period of time. When a row is powered on, the forty-eight disk drives may use 5.5 watts each maximum, just on power up, thus drawing 264 watts maximum for a short period of time. In some embodiments, two of the 240-Watt DC-to-DC power supplies are wired in parallel to provide this power requirement. Some embodiments provide additional individually activated relay switches, such that fewer disk drives (e.g., twenty-four at a time) are powered on at any one time. In some embodiments, two rows are powered on simultaneously, using different pairs of DC-to-DC power supplies. In some embodiments, a plot of disk-drive power over time at power up shows transient power to be below 0.5 amps after 3 seconds, but even if it is 10 or 15 seconds, or some other value; some embodiments provide a programmable delay between the power up of rows to keep the power draw well within the capability of the power supplies.

Sequencer timing and power control, in some embodiments, is simple, easy to develop and inexpensive. Some embodiments use one or more PIC-brand controllers (model PIC16F872, an 8-bit high-performance RISC CPU available from Microchip Technology Inc., Chandler, Ariz., is used for some embodiments) that are RISC-based CMOS technology and have an interface for chip-to-chip communication. In some embodiments, they provide temperature sensing and full environmental control. In some embodiments, the controller is made using one of the chip sets (such as model VSC7160 12-Port SAS Expander that can run at 1.5 Gbps and 3.0 Gbps, and that includes Table Routing and a Serial SCSI Protocol (SSP) engine, or model VSC7151 9-Port Serial Attached SCSI Edge Expander that can run at 1.5 Gbps and 3.0 Gbps) from Vitesse, or other suitable controller and/or expander chip sets for just-a-bunch-of-disks (JBOD) control.

FIG. 3B is a schematic of a disk-drive data-storage apparatus 204 having a power supply 300′. In some embodiments, apparatus 204 includes a first circuit board 381 and a first plurality of disk-drive connectors 311 that are operatively coupled to the first circuit board 381. The apparatus also includes a first plurality of electrically controlled relay switches 378 that include a first relay switch 320, a second relay switch 322, a third relay switch 326, and a fourth relay switch 324. The apparatus also includes a first plurality of DC-to-DC power supplies 374 that includes a first DC-to-DC power supply 312 and a second DC-to-DC power supply 314 that are operatively coupled to the first circuit board 381. In some embodiments, the DC-to-DC power supplies 374 receive an intermediate power voltage. In some embodiments, the intermediate voltage is about 48 volts. In some embodiments, the plurality of DC-to-DC power supplies 374 are connected through the first plurality of switches 378 to supply power to each one of the first plurality of disk-drive connectors 311. The plurality of DC-to-DC power supplies 374 provide crossover power to the plurality of switches 378 such that each one of the plurality of disk-drive connectors 311 is coupled through the plurality of switches 378 to each one of the first plurality of DC-to-DC power supplies 374. Dual power inputs with crossover power being directed through the plurality of switches to a plurality of disk-drive connectors provide a redundant supply of power to the plurality of disk-drive connectors.

In some embodiments, sequencer 368 is operable to control a plurality of switches in order to sequentially power up subsets of a plurality of disk drives. Use of a sequencer reduces the magnitude of power surges occurring within the apparatus. For example, in some embodiments, the apparatus includes a sequencer 368 that is operable to control a plurality of switches 378 in order to sequentially power-up subsets 352 and 354 of the first plurality of disk-drive connectors 311. In some embodiments, sequencer 368 first activates (e.g., applies power to the relay coils) only certain switches (e.g., switches 320 and 324) that supply power to one subset of the disk drives (e.g., subset 352), and at a slightly later time (e.g., 0.5 seconds to 5 seconds later, depending on the length of time that the disk drives draw extra power to spin up), sequencer 368 then activates only certain other switches (e.g., switches 322 and 326) that supply power to one other subset of the disk drives (e.g., subset 354). This reduces the maximum power surge that must be supplied by the power supplies 374 and 376 and by the AC-to-DC power sources 370 and 372). In some embodiments, sequencer 368 later activates only certain other switches (e.g., switches 328 and 332) that supply power to one other subset of the disk drives (e.g., subset 356), and still later sequencer 368 then activates only certain other switches (e.g., switches 330 and 334) that supply power to one other subset of the disk drives (e.g., subset 358). At four still later sequential times, sequencer 368 will successively activate the relay switches 336-350 to power on subgroups 360, 362, 364, and 366. By dividing the disk drives into subgroups (e.g., eight subgroups in the embodiment described above), the power surge for spin up is quite reduced.

In some embodiments, either individual power supply 312 or 314 alone can provide enough power for all of the disk-drive connectors to which it is operatively coupled. Accordingly, if a power supply 314 fails, the redundant power supply 312 is able to provide power to the plurality of disk-drive connectors and the apparatus will continue to operate. Power supplies that can be used within an apparatus of the invention can be obtained commercially (e.g., ASTEC POWER, Carlsbad, Calif. 92008). In some embodiments, each power supply will provide less than 100 mV of ripple. In some embodiments, each power supply will produce about 50 mV of ripple. Furthermore, power supplies having a variety of voltage-ranges may be used in various embodiments. In some embodiments, an AC power supply is used that has a simple rectifier and a voltage-range of, for example 18-36VDC, 36-72VDC, 90-200VDC, 180-400VDC, or the like. In some embodiments, each power supply within an apparatus is “Hot Box” swappable which enables the power supply to be removed and replaced while the apparatus is running.

In some embodiments, the apparatus includes one or more AC-to-DC power supplies or sources 370, 372 that are operable to receive AC wall power and to generate an intermediate power voltage. In some embodiments, an intermediate power voltage ranges from about 18 volts to about 36 volts. In some embodiments, an intermediate power voltage ranges from about 36 volts to about 72 volts. In some embodiments, an intermediate power voltage ranges from about 90 volts to about 200 volts. In some embodiments, the intermediate voltage is about 48 volts of direct current.

In some embodiments, the voltage output from a power supply into each one of the switches 320 to 350 is a voltage that is suitable to be used directly by a disk drive 120 that is plugged into one or more of the plurality of disk-drive connectors 126. Examples of voltages that are suitable to be used directly by a disk drive include those within a range of 5 volts plus or minus five percent (e.g., for disk drives using the industry standard 2.5-inch form factor). In some embodiments, the suitable voltage is within a range of 3.3 volts plus or minus five percent (e.g., for disk drives using the industry standard 1.8-inch form factor). In some embodiments, the suitable voltage is some other suitable voltage selected for the disk drives used.

In some embodiments, a first switch 320 is connected to couple a first DC-to-DC power supply 312 to a first subgroup (proper subset) 352 of a first plurality of disk-drive connectors 311, and the second switch 322 is connected to couple a second DC-to-DC power supply 314 to a second proper subset 354 of the first plurality of disk-drive connectors 311.

In some embodiments, an apparatus includes a third switch 326 that is connected to couple a first DC-to-DC power supply 312 to the second proper subset 354 of the first plurality of disk-drive connectors 311, and a fourth switch 324 that is connected to couple the second DC-to-DC power supply 314 to a first proper subset 352 of the first plurality of disk-drive connectors 311.

In some embodiments, an apparatus includes a fifth switch 332 that is connected to couple the first DC-to-DC power supply 312 to a third proper subset 356 of the second plurality of disk-drive connectors 313, a sixth switch 330 that is connected to couple the second DC-to-DC power supply 314 to a fourth proper subset 358 of the second plurality of disk-drive connectors 313, a seventh switch 334 that is connected to couple the first DC-to-DC power supply 312 to the fourth proper subset 358 of the second plurality of disk-drive connectors 313, and the eighth switch 328 is connected to couple the second DC-to-DC power supply 314 to a third proper subset 356 of the second plurality of disk-drive connectors 313.

In some embodiments, an apparatus includes a third DC-to-DC power supply 316. In some embodiments, an apparatus includes a fourth DC-to-DC power supply 318.

In some embodiments, an apparatus includes a ninth switch 336 that is connected to couple a third DC-to-DC power supply 316 to a fifth proper subset 360 of a third plurality of disk-drive connectors 315. In some embodiments, an apparatus includes a tenth switch that is connected to couple a fourth DC-to-DC power supply 318 to a sixth proper subset 362 of the third plurality of disk-drive connectors 315. In some embodiments, an apparatus includes an eleventh switch 342 that is connected to couple a third DC-to-DC power supply 316 to a sixth proper subset 362 of a third plurality of disk-drive connectors 315. In some embodiments, an apparatus includes a twelfth switch 340 that is connected to couple a fourth DC-to-DC power supply 318 to a fifth proper subset 360 of a third plurality of disk-drive connectors 315.

In some embodiments, an apparatus includes a thirteenth switch 348 that is connected to couple a third DC-to-DC power supply 316 to a seventh proper subset 364 of a fourth plurality of disk-drive connectors 317. In some embodiments, an apparatus includes a fourteenth switch 346 that is connected to couple a fourth DC-to-DC power supply 318 to an eighth proper subset 366 of a fourth plurality of disk-drive connectors 317. In some embodiments, an apparatus includes a fifteenth switch 350 that is connected to couple a third DC-to-DC power supply 316 to an eighth proper subset 366 of a fourth plurality of disk-drive connectors 317. In some embodiments, an apparatus includes a sixteenth switch 344 that is connected to couple a fourth DC-to-DC power supply 318 to a seventh proper subset 364 of a fourth plurality of disk-drive connectors 317.

In some embodiments, the apparatus includes a sequencer 368 that is operatively coupled to each one of the plurality of switches 378, 380, 382, and 384 and operable to apply power in a sequence over a period of time to the plurality of switches 378, 380, 382, and 384 in order to reduce the magnitude of power-on surge.

In some embodiments, the apparatus includes a second circuit board 383 to which a second plurality of disk-drive connectors 313 are each operably coupled. In some embodiments, an apparatus includes a third DC-to-DC power supply 316 and a fourth DC-to-DC power supply 318 that are both operably coupled to a second circuit board 383.

In some embodiments, the apparatus includes a plurality of disk drives connected to a first plurality of disk-drive connectors 311.

In some embodiments, the apparatus is included within an enclosure. In some embodiments, the enclosure includes a first air-inlet manifold 1112 configured to direct air between a first plurality of disk drives and a first air-outlet manifold 1114 configured to receive warmed air and direct the warmed air out of the enclosure.

In some embodiments, an apparatus includes a multiprocessor having two or more processing units and a memory coupled to the processing units, wherein the memory is operable to send and receive data from a first plurality of disk drives.

In some embodiments, an apparatus includes a video-streaming subsystem, the video-streaming subsystem including one or more processing units and a memory coupled to the one or more processing units and operable to send and receive data from the first plurality of disk drives and to simultaneously output a plurality of video streams.

In some embodiments, an apparatus includes a video-on-demand controller operable to receive requests for video programming from each one of a plurality of users, and to access and direct video output to the plurality of users based on the requests.

In some embodiments, the invention provides a method that includes operatively coupling a first plurality of disk-drive connectors 311 to a first circuit board 381, operatively coupling a first plurality of DC-to-DC power supplies 374 to the first circuit board 381, and connecting the DC-to-DC power supplies 374 through a first plurality of electrically controlled relay switches 378 to supply power to each one of the first plurality of disk-drive connectors 311. The plurality of power supplies 374 provide crossover power to the plurality of switches 378 such that each one of the plurality of disk-drive connectors 311 is coupled through the plurality of switches 378 to each one of the first plurality of DC-to-DC power supplies 374. In some embodiments, the first plurality of electrically controlled relay switches 378 includes a first switch 320 and a second switch 322. In some embodiments, the DC-to-DC power supplies 374 receive an intermediate power voltage. In some embodiments, the intermediate voltage is about 48 volts of direct current. In some embodiments, the first plurality of DC-to-DC power supplies 374 includes a first DC-to-DC power supply 312 and a second DC-to-DC power supply 314.

In some embodiments, the method includes operatively coupling a sequencer 368 to control a first plurality of switches 378 in order to sequentially power up a first proper subset 352 and a second proper subset 354 of a first plurality of disk-drive connectors 311 over a period of time.

In some embodiments, the method includes providing an AC-to-DC power supply 370 that is operable to receive AC wall power and to generate an intermediate power voltage.

In some embodiments, the method includes providing an AC-to-DC power supply 370 having an intermediate voltage that is about 48 volts of direct current. In some embodiments, the voltage output from the AC-to-DC power supply 370 into each one of the switches 378 is a voltage suitable to be directly used by a disk drive that is plugged into one or more of the plurality of disk-drive connectors 311.

In some embodiments, the method includes connecting a first switch 320 to couple a first DC-to-DC power supply 312 to a first proper subset 352 of a first plurality of disk-drive connectors 311, and connecting a second switch 322 to couple a second DC-to-DC power supply 314 to a second proper subset 354 of a first plurality of disk-drive connectors 311. In some embodiments, the method includes connecting a third switch 326 to couple a first DC-to-DC power supply 312 to a second proper subset 354 of a first plurality of disk-drive connectors 311, and connecting a fourth switch 324 to couple a second DC-to-DC power supply 314 to a first proper subset 352 of a first plurality of disk-drive connectors 311.

In some embodiments, the method includes connecting a fifth switch 332 to couple a first DC-to-DC power supply 312 to a third proper subset 356 of a second plurality of disk-drive connectors 313. In some embodiments, the method includes connecting a sixth switch 330 to couple a second DC-to-DC power supply 314 to a fourth proper subset 358 of a second plurality of disk-drive connectors. In some embodiments, the method includes connecting a seventh switch 334 to couple a first DC-to-DC power supply 312 to a fourth proper subset 358 of a second plurality of disk-drive connectors 313. In some embodiments, the method includes connecting an eighth switch 328 to couple a second DC-to-DC power supply 314 to a third proper subset 356 of a second plurality of disk-drive connectors 313.

In some embodiments, the method includes connecting a ninth switch 336 to couple a third DC-to-DC power supply 316 to a fifth proper subset 360 of a third plurality of disk-drive connectors 315. In some embodiments, the method includes connecting a tenth switch 338 to couple a fourth DC-to-DC power supply 318 to a sixth proper subset 362 of a third plurality of disk-drive connectors 315. In some embodiments, the method includes connecting an eleventh switch 342 to couple a third DC-to-DC power supply 316 to a sixth proper subset 362 of a third plurality of disk-drive connectors 315. In some embodiments, the method includes connecting a twelfth switch 340 to couple a fourth DC-to-DC power supply 318 to a fifth proper subset 360 of a third plurality of disk-drive connectors 315.

In some embodiments, the method includes connecting a thirteenth switch 348 to couple a third DC-to-DC power supply 316 to a seventh proper subset 364 of a fourth plurality of disk-drive connectors 317. In some embodiments, the method includes connecting a fourteenth switch 346 to couple a fourth DC-to-DC power supply 318 to an eighth proper subset 366 of a fourth plurality of disk-drive connectors 317. In some embodiments, the method includes connecting a fifteenth switch 350 to couple a third DC-to-DC power supply 316 to an eighth proper subset 366 of a fourth plurality of disk-drive connectors 317. In some embodiments, the method includes connecting a sixteenth switch 344 to couple a fourth DC-to-DC power supply 318 to a seventh proper subset 364 of a fourth plurality of disk-drive connectors 317.

In some embodiments, the method includes operatively coupling a sequencer 368 to each one of a plurality of switches 378, 380, 382, and 384 that are operable to apply power in a sequence over a period of time to the plurality of switches 378, 380, 382, and 384 in order to reduce the magnitude of power-on surge.

In some embodiments, the method includes operably coupling a second plurality of disk-drive connectors 313 to a second circuit board 383, and operably coupling a third DC-to-DC power supply 316 and a fourth DC-to-DC power supply 318 to the second circuit board 383.

In some embodiments, the method includes including the apparatus 300 within an enclosure. In some embodiments, the enclosure forms a first air-inlet manifold 1112 configured to direct air between a first plurality of disk drives and a first air-outlet manifold 1114 configured to receive warmed air and direct the warmed air out of the enclosure.

In some embodiments, the method includes providing a multiprocessor that includes two or more processing units and a memory coupled to the processing units and that is operable to send and receive data from a first plurality of disk drives.

In some embodiments, the method includes providing a video-streaming subsystem that includes one or more processing units and a memory coupled to the one or more processing units that are operable to send and receive data from a first plurality of disk drives and to simultaneously output a plurality of video streams.

In some embodiments, the method includes providing a video-on-demand controller operable to receive requests for video programming from each one of a plurality of users, and to access and direct video output to the plurality of users based on the requests.

FIG. 3C is a schematic of a disk-drive data-storage apparatus 204″ having a power supply 300″. In some embodiments, apparatus 204″ includes a first plurality of disk-drive connectors 311 that are operatively coupled to a circuit board. The apparatus also includes a first plurality of electrically controlled voltage regulators 312″-314″ that are controlled by power-up sequencer 368 and connected to provide redundant sources of operating voltage to disk drives 120 in the subgroup of disk drives connected to connectors 311. The apparatus also includes a second plurality of electrically controlled voltage regulators 316″-318″ that are operatively coupled to provide redundant sources of operating voltage to disk drives 120 in subgroup 315. In some embodiments, the electrically controlled voltage regulators 312″-318″ receive DC power from one of a plurality of sources 388 of an intermediate power voltage. In some embodiments, the intermediate voltage is about forty-eight volts.

FIG. 4A is a perspective drawing of disk drive 120 mounted in a perpendicular-to-the-major-face-of-the-enclosure orientation in a disk-drive system 400. This disk drive 120 is as described for FIG. 1 above.

In some embodiments (as shown in FIG. 4A), a second disk drive 120′ is mounted face-to-face, substantially parallel to, and adjacent to drive 120, such that if simultaneous seek operations are performed to both drives from the same starting position and to the same ending track, the two rotational accelerations will at least partially cancel. With respect to drive 120′ and its Z_R120′ center-of-mass axis (in some embodiments, Z_R120′ is collinear with, and in the opposite direction as, Z_R120), accelerations 147 around its Z_R120′ center-of-mass axis are in the opposite direction (clockwise versus counterclockwise) and approximately equal in magnitude as accelerations 187 of drive 120.

Rotational and translational forces that are produced by a disk drive can be transmitted to other disk drives. For example, if the front corner 119 (the corner furthest from actuator axis 111) is moved or rotated downward (as a result of torque 192) relative to the rest of the drive (and/or corner 121 is moved relatively upward), the actuator 112 will rotate in a direction 191 taking the head 114 off its track 113. Conversely, if the actuator 112 rotates in a direction 191 for its seek operation, the front corner 119 moves downward 192 relative to the rest of the drive, transmitting rotational force to other drives in its neighborhood. Moving a head off track during a read or write operation causes a loss in performance, since an entire disk revolution is needed to get back to the data that was missed when the head moved off track.

Disk drives can be arranged through use of the methods of the invention to reduce transmission of rotational forces to neighboring disk drives. Additionally, the invention provides multiple disk drives that are arranged within an apparatus so that transmission of rotational forces from one disk drive to a neighboring disk drive is reduced. In some embodiments of the invention, a second drive 120′ is placed back-to-back to drive 120, such that its disks 145 are rotating in the opposite direction as disks 115, and its actuator 142 moves in the opposite direction around its axis 141 as does actuator 112 relative to an outside frame of reference. In some embodiments, connector 116 of drive 120 is plugged into socket 126 on board 150, and is held by one or more visco-elastomeric (or, in some embodiments, elastomeric, rubbery, soft plastic or otherwise compliant to some degree) holder(s) 127 and 128. Similarly, connector 156 of drive 120′ is plugged into socket 166 on board 150, and is held by one or more visco-elastomeric (or, in some embodiments, elastomeric) holder(s) 167. In some embodiments, drives 120 and 120′ are mounted so that their Z_Rcenter-of-mass axes are aligned, and actuators 112 and 142 are driven with substantially simultaneous operations, in order to cancel some or all of the rotational force due to their respective seek operations.

In contrast to rotational forces, an up or down movement of board 150 at location 118 directly under the drive's center of rotational mass will merely cause a translation motion in the Y_Tdirection 182, which does not cause a rotation around the Z_Rcenter-of-mass axis, and thus does not cause tracking errors in drive 120. Thus, a rotational force received at point 118 causes fewer problems than if at corner 119 or corner 121 of drive 120. Further, if the actuator 112 moves in a direction 191 for its seek operation, the point 118 does not move upward or downward, but experiences a minor twist, transmitting very little rotational force to other drives if their corner 119 is closest to this point 118 on drive 120. Thus, very little rotational force is transmitted from point 118; this causes fewer problems to neighboring drives if their corner 119 or corner 121 is closest to this point 118.

Translational displacements 180 which move the entire drive 120 in X_Tdirection 181, Y_Tdirection 182, or Z_Rdirection 183 generally do not cause tracking errors, nor does rotational acceleration 185 around the X_Rcenter-of-mass axis or rotational acceleration 186 around the Y_Rcenter-of-mass axis. However, a rotational acceleration 187 around its Z_Rcenter-of-mass axis is problematic, as described above.

Disk drives are generally able to function adequately in environments that induce/transmit translational vibrations along 3 axes of the drive (translational movements along X_T, Y_Tand Z_Twill not move the head off track, since the actuator is generally quite balanced on its rotational axis) and angular acceleration or rotational force about 2 axes (X_R, Y_R; see FIG. 4A) also do not generally move the head off track. However, rotational force that is transmitted to the head-disk assembly (HDA) around the Z_R-axis is problematic. Rotation of the actuator around this axis is what moves the head that is attached to the actuator from track-to-track. When caused by the actuator motor, this moves the head to the desired track during a seek operation. However, when its neighboring drives transmit rotational force to a drive, sector-tracking problems can occur. Even a very small amount of rotational force is known to increase the position-error signal of the head, cause instability in the servo system, degrade I/O performance, increase power consumption and increase error rates of disk drives. During any seek operation, an HDA using a rotary head actuator generates rotational force in a direction opposite to that of the acceleration of the head actuator, and transmits this energy to the environment around it, including other disk drives. Disk drives are most sensitive to rotational force during the sector-tracking media transfer phase of operation, but are less sensitive to rotational force during a seek operation.

The following aspects and embodiments of the invention are aimed at reducing the effects of rotational and translational forces among a plurality of disk drives mounted in a mechanical enclosure. In addition, where RAID hardware or software logic is used to increase the performance and/or reliability of a plurality of disk drives, the following aspects and embodiments also describe how the disk drives can be arranged mechanically in relation to one another and in relation to RAID striping and mirroring logic to reduce the effects of rotational and/or translational forces.

EMBODIMENT A1

Counter-Rotating Disk Drives in a Mirrored Set to Offset Rotational Acceleration Vibration (RAV)

“Mirrored disks” are a set of M (where M is two or greater) disk drives that are logically connected as a set and at least some of the data written to that logical set is replicated to each of the M drives for each write operation. In some embodiments, all data sent to the set of drives is replicated, while in other embodiments, some amount (e.g., one-hundred-fifty GB) or some percentage of the drive's data space (e.g., fifty percent) is mirrored and the remaining data on each drive is unique or different, in order to provide mirrored speed and redundancy for the portion that is replicated, while also providing a lower cost per gigabyte for the other data by writing only a single copy. The processor elements (PEs) or operating system (OS), in some embodiments, could see a set of four three-hundred-GB drives as one four-way-mirrored drive of one-hundred-fifty GB, plus four non-mirrored drives of one-hundred-fifty GB each. In some embodiments, some portion or percentage of the data is replicated with a higher number of copies (e.g., a set of four three-hundred-GB drives could have thirty percent of the data or 90 GB replicated four times, once for each disk drive, with the operating system software seeing one 4-way-mirrored ninety-GB drive), while other data is replicated across fewer drives (e.g., ninety GB replicated twice to a first pair of drives, and another 90 GB replicated twice to a second pair of drives, so the OS sees two 2-way-mirrored ninety-GB drives), and/or split differently (e.g., one-hundred-twenty GB replicated thrice to three drives, and another one-hundred-twenty GB sent as non-mirrored data to a fourth drive, so the OS sees one 3-way-mirrored 120-GB drive and one non-mirrored one-hundred-twenty-GB drive). In such mirrored embodiments, every full-mirrored write operation is sent to all N drives, so every drive has a copy of all the data, while subset-mirrored writes are sent only to the specified subset. In some embodiments, each of a plurality of subsets of the drives have drives placed alternately back-to-back or front-to-front, as shown in FIG. 4A, so that half of the drives are rotating in the opposite direction as the other half. In some embodiments, read operations are also sent to all N drives (or to all of the subset of drives having the replicated data), so the drive that can return the data fastest has its data used, and the other drives' data is discarded. This provides the increased reliability of the duplicated data, and increases read performance to that of the drive that happens to have the least rotational latency (by the happy chance of having the rotational angle of its disks closest to the requested data) to reach the requested data. Further, since all seek operations (reads and writes) are sent to all M drives (or subset of M drives) of the set at substantially at the same time, the rotational accelerations of the M simultaneous seek operations cancel, at least to some extent. Further there are no seek operations for some of the drives while others of this set of M drives are reading or writing, tracking errors due to RAV are reduced.

EMBODIMENT A2

Counter-Rotating Disk Drives in a Mirrored Set to at Least Partially Offset RAV, Optionally Also Using Read Splitting

Again, every write operation is sent to all M drives, so every drive has a copy of all the data. In some embodiments, each read operation is sent to only one of the M drives, so the other drives have less utilization and can accept read operations to retrieve other data. This provides the increased reliability of the duplicated data, and increases read performance since more drives can be performing separate read operations simultaneously. Again, the drives are placed alternately back-to-back or front-to-front, as shown in FIG. 4A, so that half of the drives are rotating in the opposite direction as the other half. Since all write seek operations (only for writes) are sent to all M drives of the set, the rotational accelerations of the write-seek operations cancel, at least to some extent. Further, to the extent that probability allows, the read-seek operation to one drive will not occur during the read-data-tracking portion of a read to another drive of the set of M drives. Since all drives have the same data, four successive read commands to any of the data can each be sent to a different drive.

In some embodiments, read operations to large blocks of data are broken into smaller read commands, each to a different portion of the data, and each sent substantially simultaneously to a different drive of the set. Thus, if M=4, a read operation to fetch, for example, a 640-KB block of data is broken into four 160-KB read operations, each sent substantially at the same time to a different drive of the set. Thus, four seeks of substantially the same duration and to approximately the same locations on each drive will occur at about the same time. Two would have a clockwise acceleration and the other two would be counter-clockwise. The first drive would return the first 160-KB portion of the 640 KB-read request, the second drive would return the second 160-KB portion, the third drive would return the third 160-KB portion, and the fourth drive would return the fourth 160-KB portion. This provides the advantage of the counter-rotating seek commands canceling some of the RAV, the seek operations occurring when the other drives are not trying to keep on track and not occurring when heads are trying to stay on track, and the speed of parallel data retrieval providing improved performance.

Some embodiments use vulnerability mapping, described below, as one basis for selecting which drive or drives are to be used for a read-split read (i.e., a read operation that could be satisfied by data stored on any one of a mirrored set of drives since all prior write operations replicated their data on all drives of that mirrored set).

EMBODIMENT A3

Counter-Rotating Disk Alternation in a Striped Set to Offset RAV

“Striped disks” are a set of N disk drives that are logically connected as a set and data written to that logical set is spread across the set. At some level of granularity, a block of data is broken into sub-blocks, wherein each successive sub-block is written to a different drive. Thus, the block need not wait to be entirely written to or read from one drive in a serial manner, but instead the set of drives works in parallel, each writing or reading their portion of the block. The set of striped disks are, or can be, viewed by the system's processors as a single logical disk drive having a capacity that is the sum of the capacities of all drives in the set, and wherein each successive block of data (where a block can be any convenient size, such as 512 bytes, 8192 bytes, or any other desired size) is written to a different drive (with a plurality of N drives, every Nth block is written to the first drive, every N+1^stblock is written to the second drive, and so on). When data is written to or read from the logical disk that includes the set of striped physical disk drives, a single I/O request to the logical disk frequently spans two or more logically adjacent physical disk drives (each having one stripe of the data), and as a result, there is a high probability of simultaneous actuator-seek movements among these neighboring head-disk assemblies (HDAs).

In some embodiments, the minimum processor-level block size is made to be an N multiple of the minimum drive-level block size, such that every processor-level read or write is automatically striped across N drives of a set. For example, if N=4 and the processor-level block size is made 8 KB, the drive-level block size is made 2 KB, and each read operation from the processor causes a read operation to each one of the N drives. An 8-KB processor read causes four 2-KB read operations, while a 16-KB processor read causes four 4-KB read operations, one to each of the drives of the set. A 56-KB write operation causes four 14-KB write operations, one to each of the drives of the set. The N operations, one to each one of the N drives in a set will be to the same logical address on each drive, and thus cause substantially simultaneous seek accelerations that cancel if the drives are alternately clockwise and counterclockwise. Since the N operations that are sent to the N drives each access 1/N of the data, the data-transfer phase is shortened. Often, the seek accelerations are rotational accelerations that tend to be substantially simultaneous, and similar in duration, speed/acceleration, direction and frequency in the N drives of a set. By alternating the position (face-to-face or back-to-back) of each of the drives in the stripe, the combined rotational accelerations of the HDAs will, by design, offset one another other as shown in FIG. 4A. In some embodiments, for example, a 16-KB read operation from the system is broken into two 8-KB operations to logically adjacent drives in a set, where a first drive has a seek in a clockwise direction as seen from its top cover, and a second drive also has a seek in a clockwise direction as seen from its top cover, but when the top covers are face-to-face adjacent, these two rotational accelerations are in opposite rotational directions and at least partially cancel each other's mechanical motions. Coupling disk drives in this way, both mechanically and also to the RAID striping logic, t