Browse Prior Art Database

Algorithm for Determining an Optimal Backup Strategy

IP.com Disclosure Number: IPCOM000108271D
Original Publication Date: 1992-May-01
Included in the Prior Art Database: 2005-Mar-22
Document File: 4 page(s) / 124K

Publishing Venue

IBM

Related People

Schwendemann, W: AUTHOR [+2]

Abstract

Database Administrator needs a quantitative measure for determining the backup strategy for a database. The choice between a full backup or an incremental backup is not always obvious. The following algorithm can be used to determine what backup strategy should be employed. DEFINITION OF BACKUP STRATEGIES FULL BACKUP Backup entire database onto an archival storage media. When the database is being restored, it is replaced with the copy that is contained on that storage media. INCREMENTAL BACKUP Backup only those pages which have changed (i.e., been updated) since the last backup. The backup procedure would check the LSN (i.e., logical sequence number), which is essentially a time stamp, and write the page to the backup media if it is greater than the date of the last backup. DETERMINING BACKUP STRATEGY

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 50% of the total text.

Algorithm for Determining an Optimal Backup Strategy

       Database Administrator needs a quantitative measure for
determining the backup strategy for a database. The choice between a
full backup or an incremental backup is not always obvious. The
following algorithm can be used to determine what backup strategy
should be employed.
DEFINITION OF BACKUP STRATEGIES
   FULL BACKUP Backup entire database onto an archival storage media.
When the database is being restored, it is replaced with the copy
that is contained on that storage media.
   INCREMENTAL BACKUP Backup only those pages which have changed
(i.e., been updated) since the last backup. The backup procedure
would check the LSN (i.e., logical sequence number), which is
essentially a time stamp, and write the page to the backup media if
it is greater than the date of the last backup.
DETERMINING BACKUP STRATEGY

      There are many factors to consider when deciding upon on backup
method (i.e., full or incremental). The most important of these are:
   o   CPU and I/O speed
   o   volatility of database
   o   duration between backups
   o   costs associated with restoration of database
VARIABLE DEFINITIONS
   VARIABLE                      MEANING
   COST(FULL BACKUP)             cost of a full backup
   IO(READ)                      I/O cost of reading a database page.
   IO(WRITE)                     I/O cost of writing a database page
to backup media.
   COST(INCREMENTAL BACKUP)      cost of incremental backup.
   CPU(STATUS)                   CPU cost associated with checking
the status of a database page in order to determine if it has been
modified since the last backup.
   N                             the total number of pages the data
base is comprised of.
   M                             percentage of database pages that
have been modified.
ASSUMPTIONS
   o   The cost of determining whether a page needs to be written to
the backup media (i.e., CPU [status]) is linear proportional to the
total number of database pages.
   o   The I/O costs associated with writing a database page to the
backup media (i.e., IO[write]) is linear proportional to the number
of pages transferred.
   o   The ratio of changed/total database pages (i.e., variable m)
can be obtained at run time. A counter would be incremented whenever
a page was updated for the first time since the last backup.  The
counter would be reset to zero after every backup. A count of how
many pages comprised the database would also be kept.
   o   The speed of an I/O device can be obtained from the hardware
documentation.
   o   The I/O cost of reading a database, assuming a 4k page size,
is the time needed for a page fault to complete.
   o   The cost of checking the page status (i.e., CPU (status)) can
be d...