Browse Prior Art Database

Using Dual Actuator Shared Data Direct Access Storage Devices Drives In a Redundant Array

IP.com Disclosure Number: IPCOM000119312D
Original Publication Date: 1991-Jan-01
Included in the Prior Art Database: 2005-Apr-01
Document File: 3 page(s) / 138K

Publishing Venue

IBM

Related People

Floryance, GG: AUTHOR

Abstract

Described is a technique for using Dual Actuator Shared Data Direct Access Storage Devices (DASDs) in a disk array to improve performance while also providing the additional advantage of reducing cost by eliminating the need for extended battery backup.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 49% of the total text.

Using Dual Actuator Shared Data Direct Access Storage Devices Drives
In a Redundant Array

      Described is a technique for using Dual Actuator Shared
Data Direct Access Storage Devices (DASDs) in a disk array to improve
performance while also providing the additional advantage of reducing
cost by eliminating the need for extended battery backup.

      A dual actuator shared data DASD is a device which has two
separate actuators accessing the same data surface of a single
spindle.  Each actuator has dedicated read/write electronics allowing
it to operate independently from the other actuator.   Data read or
written by one actuator can be read or written by the second
actuator.

      The host may view the DASD subsystem as a single logical
device.  The DASD subsystem can be software configured to implement
one or more array techniques including mirroring for transactional
processing, parallel disk arrays for high data transfer rates and
journaling of DASD writes.  The DASD subsystem analyzes data usage
and moves the data between the various sub-regions (i.e., from a
mirrored to a parallel area) optimizing performance and redundancy.
The dual actuator shared data drive enhances the hardware platform
and provides increased flexibility for software configuration to meet
the performance and redundancy requirements of the specific array
application.

      The concept of this invention is illustrated in the figure.
The disk drives are shown as circles with each attached rectangular
box representing Read/Write electronics.  Disk drives M1 and M2 are a
mirrored pair with dual actuators sharing the same data.  Disk drives
P1 through Pn are drives in the parallel array implementing either
Striped ChkSum, Parity or some other form of array technology.   The
arrows on each bus indicate the direction of data flow.

      To reduce or eliminate the need for battery backup, a pair of
mirrored drives (M1,M2) are used for most write operations.  (The
DASD Subsystem may determine that the data should be written directly
to the Parallel Array.) Mirroring the drives ensures that the loss of
one drive will not cause a loss of data in the subsystem.  Dual
actuator drives also improve performance by use as a fast write
buffer.

      During normal operation, the BUSIN channel of drives M1,M2 will
transfer data to the media as it is received from the host.  This can
be done without regard to the logical block address of the data.
This is, in effect, a journaling procedure which writes data to the
media in a continuous fashion and allows the media to continue to
spiral in (via track seeks) without taking a latency or seek time
delay. As the data is written to the media, a directory is maintained
in the controller which records the location and content of each
block of data written.  This directory can be used to reconstruct the
data base in the event that the device was suddenly powered down.  A
result of the journaling proce...