Electrical computers and digital processing systems: memory – Storage accessing and control – Specific memory composition
Utility Patent
1999-03-26
2001-01-02
Yoo, Do Hyun (Department: 2759)
Electrical computers and digital processing systems: memory
Storage accessing and control
Specific memory composition
C711S004000, C711S104000, C711S112000, C711S161000, C711S162000, C711S165000, C360S008000, C360S015000, C360S047000, C369S084000
Utility Patent
active
06170037
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to disc drive storage subsystems.
DISCUSSION OF THE RELATED ART
Disc drives are used in many different types of computer or data processing systems to store data. Disc drives include one or more discs of a recording medium (e.g., a magnetic recording medium or an optical recording medium) on which information can be written for storage purposes, and from which stored information can be read. The recording medium is typically in the form of a disc
1
that rotates about a spindle
2
as shown in FIG.
1
. The disc includes a plurality of tracks on which information is recorded. In
FIG. 1
, only an outer track
3
and an inner track
5
are shown to simplify the drawing. However, it should be appreciated that a surface of an actual recording disc will include a large number of tracks (e.g., a 9 GB drive includes over 5,000 tracks) in addition to the outer and inner tracks
3
,
5
. In a disc drive that includes multiple discs, the discs are conventionally stacked so that corresponding tracks overlie one another.
As shown in
FIG. 1
, each of the tracks is conventionally subdivided into a plurality of sectors
7
(also known as blocks). The sectors
7
define the smallest amount of data that is written to or read from the disc in one operation. An exemplary size for each sector is 512 bytes, which is the standard for disc drives that communicate with other components of a data processing system over a SCSI interface.
Data is read from and written to the disc
1
using a head
9
that is positioned adjacent (e.g., above) the surface of the disc via an arm
11
. In operation, the disc is rotated at a high rate of speed (e.g., 5,400 rpm, 7,200 rpm or 10,033 rpm). The arm
11
is pivoted by an actuator (not shown) about a pivot point
12
to move the head
9
in a seek direction (indicated by the arrow S in
FIG. 1
) so that the head can be positioned above any of the tracks
3
,
5
of the disc. The combination of the rotation of the disc and the movement of the head
9
in the seek direction S enables the head to be positioned adjacent any sector
7
of the disc to access (i.e., read information from or write information to) that sector.
The performance of a disc drive system is largely impacted by three system characteristics, i.e., seek time, latency and data rate. The seek time relates to the delay incurred in positioning the head
9
above the appropriate track. In the worst case, the seek time is defined by the delay incurred in moving the head
9
between the inner and outer tracks
5
,
3
. The latency of the system is the time it takes for the rotation of the disc
1
to bring the desired sector
7
to a position underlying the head
9
. The worst case latency is defined by the time it takes to complete a single rotation of the disc. Finally, the data rate of the system relates to how quickly data can be read from or written to the disc once the head
9
is positioned above the appropriate sector
7
. The data rate is dependent upon the bit density of the information stored on the disc, the rate of rotation of the disc and the disc drive electronics that process the data.
Most conventional disc drive systems attempt to maximize the storage capacity of the disc
1
. Thus, the disc
1
is typically provided with as many tracks as possible, and each track is provided with as many sectors as possible. Although maximizing storage capacity, such configurations result in limitations being placed on the performance of the system. For example, maximizing the number of tracks on the disc results in a long worst case seek time, because the head
9
must move across substantially the entire radius of the disc.
In addition, the relative performance of the disc drive system is greater when accessing tracks that are positioned nearer the outer surface of the disc (e.g., track
3
) than the center of the disc (e.g., track
5
). Many disc drive systems employ a technique known as zoned constant velocity in which the total disc capacity is increased by varying the number of sectors per track with the distance of the track from the center of the disc. This technique is also called zone bit recording. A drive that employs this technique is known as a notched drive according to the SCSI specification. The tracks are typically grouped into zones with each track in a zone including the same number of sectors. Outer tracks have more sectors than inner tracks. As a result, when the disc rotates, the rate of data passing by the head when accessing the outer track
3
is significantly greater than when accessing the inner track
5
, because the outer track moves past the head at a significantly faster speed and has more sectors per track. Thus, the data rate of the system is greater when reading a track positioned near the outer surface of the disc. In addition, since the outer track
3
may have relatively more information stored thereon, less seeking between tracks is required when accessing the outer tracks.
As a result, when the entire surface of the disc is used to support tracks in a conventional implementation as shown in
FIG. 1
, the performance of the disc drive system is limited by the relatively poorer performance of the system when accessing the inner tracks.
Some storage subsystems that include a plurality of disc drives have employed techniques for distributing logical volumes of information across the multiple disc drives to increase performance of the storage subsystem. One such technique is known as striping, which involves the writing of contiguous groups of blocks to different discs.
FIG. 2
illustrates a storage subsystem
15
that implements striping across two disc drives
17
and
19
(also referenced in
FIG. 2
respectively as DRIVE_
0
and DRIVE_
1
). In
FIG. 2
, the two disc drives are shown as storing eight groups of blocks that are logically contiguous in succession from G
1
through G
8
. However, it should be appreciated that actual disc drives may store significantly more groups of blocks. As shown in
FIG. 2
, in accordance with a conventional striping technique, pairs of contiguous block groups are stored in different disc drives (e.g., G
2
is stored in disc drive
19
, while G
1
and G
3
are stored in disc drive
17
). This type of striping technique is particularly advantageous when the block groups are accessed sequentially. In particular, reading operations for block groups G
1
and G
2
can be performed in parallel, because they are performed on different disc drives. Thus, when accessing a sequence of contiguous block groups, a storage subsystem that employs a striping technique can achieve higher performance than if the sequence of blocks was stored on a single disc drive.
Although striping can provide performance improvements for a storage subsystem, particularly when accessing sequential information, this technique has a number of disadvantages. First, to increase the performance of a storage subsystem having multiple disc surfaces for storing blocks of information, additional disc surfaces must be added. To add additional disc surfaces to a striped system, substantially all of the data must be moved and reallocated across all of the disc surfaces, which can be a time consuming process. This makes a striped system particularly inflexible for responding to changing performance requirements because significant costs are associated with re-configuring the striped system to increase performance.
Another disadvantage with conventional striping techniques is the inability to effectively handle hot spots, which are areas in a disc drive or storage subsystem that are accessed with great frequency. For example, referring to the striped system in
FIG. 2
, if block groups G
2
and G
4
in DRIVE_
1
were accessed consecutively with great frequency, they would need to be accessed sequentially by the disc drive
19
, which would not enable any performance improvements to be recognized due to the use of two disc drives in the subsystem. The only way to attempt to alleviate this problem in a striped system is by r
EMC Corporation
Nguyen Than
Wolf Greenfield & Sacks P.C.
Yoo Do Hyun
LandOfFree
Method and apparatus for storing information among a... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for storing information among a..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for storing information among a... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2474864