Electrical computers and digital processing systems: memory – Storage accessing and control – Control technique
Reexamination Certificate
1998-05-22
2001-02-13
Cabeca, John W. (Department: 2751)
Electrical computers and digital processing systems: memory
Storage accessing and control
Control technique
C711S161000, C711S005000
Reexamination Certificate
active
06189079
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a system for maintaining two copies of data at two different locations.
2. Description of the Related Art
A key feature of any disaster recovery system is the backup of critical data and updates to such backed-up data as soon as possible. The simplest method of backing up data is with off-site backups. However, the extent of successful data recovery in the event of failure is dependent upon the frequency of updates to the backed-up data. If a failure occurs between updates, then data updated but not backed-up may be lost. Thus, the greater time delay between back-ups, the less data is recoverable. On the other hand, increasing the frequency of back-ups causes delays in responding to requests to the volumes being backed-up. Volumes involved in back-up operations are typically unavailable for use by application programs. Thus, increasing the frequency of updates to the data back-up increases the time during which particular volumes involved in the back-up operations are unavailable.
One data back-up system is the Peer-to-Peer Remote Copy (PPRC) function offered by International Business Machines, Corporation (“IBM”), the assignee of the subject patent application. PPRC provides synchronous copying of data from a primary controller to a remote (secondary) controller. The PPRC system includes two controllers and two direct access storage devices (DASDs), a primary controller and associated primary DASD at one site and a secondary controller and secondary DASD at another site. Often, both controllers include a non-volatile storage unit (“NVS”), such as a battery backed up storage unit, to maintain data in the event of a power or system failure. With PPRC, when data is written to the primary NVS, the data is then transferred to the NVS in the secondary controller. At later points in time, the data in the primary and secondary NVSs will be destaged to the attached storage devices. With PPRC, there is minimal risk of losing data upon the occurrence of a failure because data is backed up at the secondary volumes synchronously with updates to the data at the primary DASD managed by the primary controller. The PPRC feature and other backup copy operations are further described in IBM publication “3990/9390 Storage Control Introduction,” document no. GA32-0098-08 (IBM Copyright 1987, 1996), which publication is incorporated herein by reference in its entirety.
One problem with prior art PPRC type back-up systems is that the tracks or volumes being backed up between the primary and secondary DASDs are unavailable for use while PPRC back-ups are being serviced. The PPRC operation does not consider the transfer of data to the secondary site complete until all the data updated at the primary DASD has been copied and backed-up to the secondary DASD. Thus, updates to the primary DASD delay response time to user requests to the volumes involved in the update because a synchronous update must be made to the secondary DASD before the volume involved is available. Response delays can occur with respect to requests to both the primary and secondary DASDs. Users making requests to a volume in either the primary or secondary DASD subject to a back-up operation must wait until the completion of the back-up before they can access the updated data.
As the data transfer size and distance between the primary and secondary DASDs increase, the service time of synchronous PPRC updates to both the primary and secondary DASDs increases. Increase in service time for data back-ups resulting from increased distances or larger data transfer sizes likewise delays the response to requests from host systems or application programs for volumes involved in PPRC updates. Moreover, the response time will also increase with respect to requests to volumes in the primary DASD not involved in PPRC updates to the secondary controller. This degradation in response time is often substantial, ranging anywhere from 10% to 200%.
SUMMARY OF THE PREFERRED EMBODIMENTS
To overcome the limitations in the prior art described above, the preferred embodiments disclose a system for copying data sets from a first storage device to a second storage device. The first storage device is managed by a first controller and the second storage device is managed by a second controller. The first controller receives a command to copy a plurality of data sets from the first storage device to the second storage device. First and second data structures are generated to include fields corresponding to the data sets. The fields initially indicate that the data sets have not been copied. For the data sets subject to the copy command, the first controller transfers a copy of the data set to the second controller. The second controller stores the copy of the data set received from the first controller in the second storage device. The field in the first data structure corresponding to the data set copied to the second storage device is then modified to indicate that the data set was copied to the second storage device. The field in the second data structure corresponding to the data set the second controller stores in the second storage device is also modified to indicate that the data set was stored in the second storage device.
In further embodiments, a host request to the second storage device for a data set that has not yet been copied over from the first storage device will cause the second controller to generate and transmit a message to the first controller requesting the requested data set. In response, the first controller accesses the data set requested by the second controller in the message and transfers the requested data set to the second controller. The second controller then stores the requested data set received from the first controller in the second storage device and transmits the requested data set to the host system.
In yet further embodiments, an update to a data set in the first storage device that has not yet been copied over to the second storage device will cause the first controller to transfer the data set to update to the second controller. The second controller will then store the data set received from the first controller in the second storage device. The first controller stores the update to the data set in the first storage device after transferring the data set prior to the update to the second controller.
The preferred embodiments allow for immediate copying of updates to data sets at a primary storage device to a secondary storage device which may be at a remote location. This immediate transfer insures that the backup of the secondary storage reflects the data sets at the time the backup request was initiated and the ability to recover data in the event of a failure at the primary storage device. Moreover, requests for data at the secondary storage device that shadows data maintained in the primary storage device are handled in a manner to insure that the current data maintained at the primary storage device is provided in response to the request. This servicing of requests is handled prior to the completion of the transfer of all data sets to the secondary storage device to improve the response time to data requests.
Another benefit of certain embodiments is that before data is updated to the primary storage device after a copy operation has been initiated, the pre-update version of the data set is transferred to the secondary storage device to insure that the secondary storage device maintains the data in existence at the time the copy operation was initiated. Such handling of updates is handled prior to the completion of the transfer of all data sets to secondary storage to insure that all updates are made.
The preferred embodiments, thus, provide a mechanism for immediately backing-up data at the primary and secondary sites and at the same time efficiently handling requests for access and updates of data before the transfer of all data has been completed.
REFERENCES:
patent: 5175849 (1992-12-01), Schneider
patent: 5404502 (1995-0
Micka William Frank
Novick Yoram
Cabeca John W.
International Business Machines - Corporation
Konrad Raynes & Victor LLP
Peugh Brian R.
Victor David W.
LandOfFree
Data copy between peer-to-peer controllers does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Data copy between peer-to-peer controllers, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Data copy between peer-to-peer controllers will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2563406