Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Remote Dual Copy of Count-Key-Data Formatted Data

IP.com Disclosure Number: IPCOM000115846D
Original Publication Date: 1995-Jun-01
Included in the Prior Art Database: 2005-Mar-30
Document File: 2 page(s) / 97K

Publishing Venue

IBM

Related People

Ouchi, NK: AUTHOR

Abstract

Count-Key-Data (CKD) data format requires that the Control Unit (CU) be synchronous with the disk drive. Extended CKD permits the processor to be non-synchronous but the CU must still synchronous. Thus, because of speed of light limited cable delays, the CU must be close to the disk drive. CKD also requires an associative addressing mechanism with time dependent searching of the recorded control data prior to a read or write operation. These time dependent data operations are compounded by error recovery that is also time dependent. In addition to data integrity requirements, the operations must be execution time competitive.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 53% of the total text.

Remote Dual Copy of Count-Key-Data Formatted Data

      Count-Key-Data (CKD) data format requires that the Control Unit
(CU) be synchronous with the disk drive.  Extended CKD permits the
processor to be non-synchronous but the CU must still synchronous.
Thus, because of speed of light limited cable delays, the CU must be
close to the disk drive.  CKD also requires an associative addressing
mechanism with time dependent searching of the recorded control data
prior to a read or write operation.  These time dependent data
operations are compounded by error recovery that is also time
dependent.  In addition to data integrity requirements, the
operations must be execution time competitive.

      Disclosed is a mechanism that uses the CKD to fixed block
transformation.  All operations are executed in CKD; all data are
stored in fixed block.  For most operations, the execution times are
comparable to native CKD devices.  The fixed block format permits
ease of use of a DASD array for high data availability.  In all
operations, the data are staged into a buffer for CKD operations.  If
data are written, the CKD data are transformed to fixed block and
stored on the DASD units.  The fixed block addressing mechanisms are
simple and avoid the need for close proximity of the data source.

      Remote copy is accomplished by transmitting the fixed block
data that are to be written into the local DASD to a remote DASD
unit.  The remote unit writes the data in the corresponding addresses
of the fixed block devices and signals a completion to the
originating DASD unit.  This is illustrated in the Figure.  The
originating DASD unit holds the data in its buffer until the copy is
complete.  If an error is encountered, the data may be retransmitted.
The performance of these operations is paced by the data rate of the
transmission link between the two DASD units.  If a high data rate
link is used, the performance of remote dual copy can be comparable
to the local DASD unit.  The system attached to the remote DASD unit
accesses the data in CKD format just as the originating unit.

Additional functions:
  1.  The remote DASD unit can dual copy or secure the transmitted
data
       and can signal completion to the originating DASD unit before
       completion of writing to DASD.
  2.  In a DASD unit with data compaction or Log Structured File, the
       data can be transmitted to a remote DASD unit.  The remote
DASD
       unit will use the same addressing structure as the sending
DASD,
    ...