Browse Prior Art Database

Concurrent Copy

IP.com Disclosure Number: IPCOM000112142D
Original Publication Date: 1994-Apr-01
Included in the Prior Art Database: 2005-Mar-26
Document File: 4 page(s) / 110K

Publishing Venue

IBM

Related People

Cohn, O: AUTHOR [+8]

Abstract

One of the major problems to providing continuous availability of computing systems is the service disruption while backing-up data. Backing-up individual data sets or entire DASD volumes requires that the data be made unavailable for updates for the duration of the backup operation.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Concurrent Copy

      One of the major problems to providing continuous availability
of computing systems is the service disruption while backing-up data.
Backing-up individual data sets or entire DASD volumes requires that
the data be made unavailable for updates for the duration of the
backup operation.

      This invention provides a mechanism to allow a copy of a set of
volumes or datasets to be taken while normal activity against them
continues.  The mechanism is general and supports Data Base
Management Systems and most data set organizations.  It arose out of
a need by customers to have a mechanism to backup copies of data
without suffering loss of access to the data during the backup.  The
primary objective of the Concurrent Copy architecture is to reduce
the time that customer data is unavailable during dump and backup
operations.

      Prior to Concurrent Copy, backup or dump operations, performed
on a data set or volume basis, were strictly under control of
software running in the host environment.  To dump the data contained
on a given volume at a specific point in time, the volume must be
protected from modifications until the dump operation is complete.
This restriction results in the customer having to terminate or
suspend other applications that use the volume until the dump or
backup has been completed.

      This architecture provides an enhancement to customer
operations by allowing the volume to remain available for write (and
read) activity while the dump or copy is being performed.  The
application that uses the volume does not have to be terminated as
the combination of host software and Storage Controller microcode
will manage the write updates that occur after the start of the dump
and preserve the integrity of the data.

      Concurrent Copy is a copy operation that the customer initiates
relative to a specific starting point in time.  The resultant copy
will reflect the data exactly as it was at initiation without any
updates that may have been made to the volume while the copy
operation was in progress.

      Write updates to the data are allowed on the tracks that are
participating in the copy.  Those writes that occur during the copy
operation will be handled in a fashion such that the target volume
does not receive the updates until the original version of the data
has been preserved.  The primary volume can be updated immediately
providing the data area that is to be modified has been preserved in
its original form due to prior activity.

      The Concurrent Copy process is realized by a combination of
host software and Storage Controller microcode.  Refer to Figure 1.
The host first informs the Storage Controller, using normal I/O
acc...