Electrical computers and digital processing systems: memory – Storage accessing and control – Specific memory composition
Reexamination Certificate
2001-01-29
2004-01-20
Padmanabhan, Mano (Department: 2188)
Electrical computers and digital processing systems: memory
Storage accessing and control
Specific memory composition
C707S793000, C713S001000, C713S002000, C713S100000
Reexamination Certificate
active
06681290
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention generally relates to computer systems, specifically to data storage devices for computer systems, and more particularly to a method of writing data to and reading data from a redundant array of independent disks (RAID) so as to substantially reduce or minimize the time required to complete the seek operations associated with such accesses.
2. Description of Related Art
Computer systems use a wide variety of devices for permanently storing data, i.e., in a non-volatile manner such that power to the computer system may be turned off but the data (including both user information and computer programs) are retained for future access. These direct access storage devices (DASDs) typically use a magnetic or optical medium to preserve the data. The most common data storage device has one or more generally circular disks or platters formed from a non-magnetic substrate, with a ferromagnetic coating. The disks are mounted on a spindle motor and rotate, and a pivoting arm assembly having electromagnetic transducers is used to read from, and write to, the disks. The media-bearing surfaces of the disks are partitioned into concentric tracks or cylinders. This particular magnetic storage device is commonly referred to as a hard disk drive (HDD), and is usually packaged in a modular enclosure so that it may be easily installed in and removed from the computer system.
Many computer systems use multiple HDDs for greater storage capability, or for fault tolerance. For example, U.S. Pat. No. 5,778,252 discloses an interface for concurrent accesses to multiple disk drives of a computer. That interface is an enhanced version of the industry-standard integrated drive electronics (IDE) disk drive. The device gangs multiple IDE drives together to present the image of a single, larger and faster IDE drive.
A more common example of a multi-disk storage system is the so-called redundant array of independent (or inexpensive) disks (RAID). In the RAID design, data is placed on multiple hard disks to allow input/output (I/O) operations to be completed in an overlapping manner, thereby improving system performance. The same data is recorded on more than one disk, so the fault tolerance of the system increases with the redundancy factor. A general example of a RAID design may be found in U.S. Pat. No. 5,592,648.
FIG. 1
illustrates, conceptually, the mapping of logical storage to physical storage in a typical RAID system
1
. A logical storage location
2
of a logical disk device
3
corresponds to two or more physical storage locations
4
,
5
on respective physical disk drives
6
,
7
. The RAID controller card maps the logical read/write commands to the physical memory arrays. Read performance is improved since multiple read operations can simultaneously execute on separate physical devices, although write performance is the same as for single disk storage.
FIG. 1
illustrates an exemplary RAID-1 system whose logical disk contains 20,000,000 logical blocks, each 512 bytes, or approximately 10 gigabytes (GB) of total memory. The logical disk is implemented physically with the two 20,000,000 block drives
6
,
7
.
In the simplest RAID design (RAID-1), the physical locations are the same (disk mirroring), which is illustrated in FIG.
1
. Other RAID systems can use various types of “striping” to partition the space of each drive
6
,
7
. The space is divided up into various sized units ranging from a typical sector of 512 bytes, to several megabytes (MB). In the RAID-10 design, the system has logical stripes in which each stripe is a RAID-1 array of drives. RAID systems can also incorporate other known features for permanent storage devices, such as error detection (e.g., parity information), error checking and correcting (ECC), or caching.
One important factor in the performance of a disk drive is the time it takes to move the transducer (read/write head) on the pivoting arm assembly from one track or cylinder to another (a “seek”). The arm assembly may have to move the entire length of the physical drive in some cases, since the entire length of a single device is used to store all primary copies of the recorded values (and all secondary copies of the values are likewise stored across the entire length of a single device). The geometry or physical layout of the disk drive's storage tracks affects seek time. Logical block numbers are assigned sequentially from the inner cylinders to the outer cylinders of the physical drive. In general, for a RAID-1 system such as shown in
FIG. 1
, the average seek distance is one-third of the drive.
In spite of the performance benefits achieved using RAID systems, seek time continues to be one of the primary causes for data access latency. It would, therefore, be desirable to devise an improved method of accessing a RAID system in order to reduce seek time. It would be further advantageous if the method could take advantage of existing hardware designs and storage geometry, so as not to add further complexity and cost to the design.
SUMMARY OF THE INVENTION
It is therefore one object of the present invention to provide an improved data storage device for a computer system.
It is another object of the present invention to provide a RAID-type data storage system (particularly a RAID-1 or RAID-10 system) having a substantially reduced average seek time.
It is yet another object of the present invention to provide a method of accessing a disk drive in a RAID-type system which takes advantage of the driver's physical storage geometry.
The foregoing objects are achieved in a data storage subsystem, generally comprising a plurality of physical data storage devices (such as hard disk drives) and means, connected to the storage devices, for defining a logical memory space wherein (i) a first portion of the logical memory space is defined to correspond to a primary storage space of a first one of the physical data storage devices, and correspond to a secondary storage space of a second one of the physical data storage devices, and (ii) a second portion of the logical memory space is defined to correspond to a primary storage space of the second physical data storage device, and correspond to a secondary storage space of the first physical data storage device. In this manner, the average seek distance for both devices is reduced, since the pivoting arm assembly which scans the surfaces of the hard disks need not travel as far to reach each physical storage address within a primary storage area.
The invention may additionally take advantage of geometric or other aspects of the storage devices which result in some portion of the device having a faster access time than another portion. More specifically, for hard disk drives which assign a larger number of logical blocks to the outer tracks of the disk than to the inner tracks, the memory controller can map the primary storage space of a given one of the drives to the outer tracks of the drive.
The above as well as additional objectives, features, and advantages of the present invention will become apparent in the following detailed written description.
REFERENCES:
patent: 5457791 (1995-10-01), Matsumoto et al.
patent: 5592648 (1997-01-01), Schultz et al.
patent: 5708793 (1998-01-01), Franaszek et al.
patent: 5778252 (1998-07-01), Sangveraphunsiri et al.
patent: 5809300 (1998-09-01), Utsumi et al.
patent: 5900007 (1999-05-01), Nunnelley et al.
patent: 5909691 (1999-06-01), Schultz et al.
patent: 6493729 (2002-12-01), Gusler et al.
Brower, Jr. William John
Jones Carl Evan
Smothers Joe Edward
Bracewell & Patterson L.L.P.
International Business Machines - Corporation
Namazi Mehdi
Padmanabhan Mano
Sullivan Robert M.
LandOfFree
Physical data layout to reduce seeks in a raid system does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Physical data layout to reduce seeks in a raid system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Physical data layout to reduce seeks in a raid system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3257490