Backup is the process of making a copy of data in case the original is lost or damaged.
Recovery is the opposite of the backup process where the backup copy of the data is restored and placed back into the system.
Backups provide a snapshot copy of data at particular points in time. Each backup copy allows the system’s data to be recovered back to the state it was in at the precise time the backup copy was made. In the event of total system failure, such as a hard disk crash or a fire that destroys the data completely, it is important to be able to recover to a point as close as possible to the time the failure occurred.
Backup and recovery protects against each of the following:
• Hardware errors and failure.
• Software errors.
• Physical theft or destruction of hardware.
• Unauthorised or unwanted changes to data, due to viruses or hacking, for example.
• Intentional changes to data that for various reasons need to be reversed.
Summary
Organisations have become very dependent on their TPSs. For example, making a reservation when the computerised reservation system is down is often impossible.
A breakdown in the TPS may stop the business. Well-designed backup and recovery procedures minimise disruptions when the TPS goes down.
A backup is another copy of the data that could be used to rebuild the system.
If the system goes down, the recovery process rebuilds the system.
The success of backup and recovery depends on implementing appropriate procedures.
Backups are usually stored offsite or stored onsite in a fireproof safe.
BackUp Procedures
FULL AND PARTIAL BACKUPS
There are three different types of backup that are commonly used within most backup
procedures –
Full backups,
Incremental backups and
Differential backups.
Both incremental and differential backups involve making partial backups.
Full Backup
As the name suggests, a full backup is a complete copy of all data within the system.
This can be a complete image of the entire hard disk(s), including the operating system, program files, configuration settings and of course data.
For most transaction processing systems it is the data that is of particular value – the software and configuration settings rarely change and are far easier to restore.
Therefore most businesses perform full backups of all their data files on a regular basis.
Full backups are the easiest to restore should failure occur. The full backup is simply copied back into the operational system. Unfortunately copying all files takes a long time and requires large amounts of storage; therefore it is often impractical to perform full backups on a daily basis. Common backup procedures specify that full backups be made on a weekly basis, usually commencing on Friday afternoons and for large systems continuing over the weekend.
Most operating systems store an archive bit along with each file. The archive bit is set to true when changes are made to a file and when a file is first created. When full backups are made all archive bits are set to false indicating that a backup copy of each file has just been made.
Incremental Backup
Incremental backups include making partial backups that copy all files that have changed or been created since the last backup–the last backup may have been a full, or partial backup.
An incremental backup therefore includes only those files where the archive bit is true. As a consequence performing an incremental backup is significantly faster and requires significantly less storage compared to a full backup.
After each file has been copied its archive bit is set to false.
Therefore if incremental backups are performed each afternoon then each incremental backup copies only those files that have been altered or created since the previous afternoon’s backup was made.
Before incremental backups can be made it is necessary to first perform a full backup.
Commonly faster incremental backups are then made on at least a daily basis.
The significant saving in backup time is counteracted by the extra time required to recover the data.
During a recover the
latest full backup is first restored, then
each partial backup is restored in the order in which they were made.
Hence files that have changed since the full backup progressively overwrite the older versions as each partial backup is restored.
Differential Backup
A differential backup uses partial backups to makes copies of all files that have been altered or created since the last full backup. If such partial backups are made each day then each will contain copies of all files within all previous partial backups since the last full backup was made.
To restore to the most recent backup requires first restoring the full backup and then restoring just the most recent partial backup.
In terms of archive bits, differential backups copy all files where the archive bit is true, however differential backups do not alter any archive bits.
Therefore over time one would expect more and more archive bits to be true and hence more and more files are included within subsequent differential backups.
The size of the differential backup continues to grow until the next full backup is completed.
Grandfather-Father-Son
Grandfather-Father-Son (GFS) Data Retention
This is a very common scheme used as a hierarchical data retention strategy.
For example, three sets of backups, such as weekly, monthly and yearly are defined. You can vary this frequency to suit the requirements of your environment.
The weekly or Son backups are rotated on a weekly basis with one graduating to Father status each month.
The monthly or Father backups are rotated on a yearly basis with one graduating to Grandfather status each year.
One or more of the graduated (Grandfather) backups is preserved for disaster recovery and archival purposes.
BackUp Procedures (Grandfather-Father-Son)
YouTube Video
Incremental vs. Differential Backup
Incremental vs. Differential Backup
TRANSACTION LOGS, MIRRORING AND ROLLBACK
Recall that transaction logs contain historical details of each transaction performed
including details of transactions that are currently being processed.
These details can be used to restore a transaction processing system back to a consistent state at some precise point in time – completed transactions can be recommitted or rolled back and
incomplete transactions can be continued or rolled back.
Incremental Forever / Progressive Backups
An incremental forever backup strategy minimizes backup windows while providing faster recovery of your data.
Rather than scheduling weekly (periodic) full backups, this backup solution requires only one initial full backup. Afterward, an ongoing (forever) sequence of incremental backups occurs. The incremental forever backup solution provides these advantages:
Reduces the amount of data that goes across the network.
Reduces data growth because all incremental backups contain only the blocks that changed since the previous backup.
No comparison with the backup target is needed since only changed blocks are identified.
Minimizes impact to the client system.
Reduces the length of the backup window.
No need to schedule an initial full backup as a separate schedule: the first issue of an incremental forever backup automatically defaults to an incremental forever full backup.
In addition, the restore process is optimized, as only the latest versions of blocks that belong to a restored backup are restored. Since the same area on the production disk is recovered only one time, the same block is not written to multiple times as implemented in prior versions of Data Protection for VMware. As a result of these advantages, incremental forever is the preferred backup strategy. However, if your backup solution requires long-term storage or tape retention of your data, consider the tasks described in Periodic full backup strategy.
(i) Discuss the advantages of a ‘grandfather, father, son’ backup procedure. (2 marks)
Question 2
One of Australia’s largest organisations recently conducted the country’s biggest e-voting project, allowing its 2 million members to vote via the internet.
Each voter was sent a personal identification number (PIN) by post. The voter logged onto a secure website with his/her membership number and PIN before the voting deadline. The e-voting system checked that the user had completed all voting details correctly according to voting rules. Once the vote was accepted, the voter’s details were ‘tagged’ to prevent multiple voting, and a confirmation was generated for the voter.
(i) Outline two backup procedures suitable for this e-voting transaction processing system so that no votes will be lost in the event of a system failure. (3 marks)
Do not use Grandfather, Father, Son methodology for this answer.
