Data Backup and Disaster Recovery

Data backup refers to the process of archiving or copying information saved in files and folders so that the information can be retrieved in case of a data loss event. “Data loss may happen due to different causes such as fire, theft, computer viruses and hardware failures among others” Lwin & Thein, 2009 p. 203. Disaster recovery, on the other hand, refers to the procedures and policies set up in an organization that focus on the recovery of vital information lost in the organization due to natural or human-induced causes. Disaster recovery focuses on Information Technology systems in an organization to ensure normal continuous operations after a disruption in operational activities.

Data Backups

There are 4 traditional backup strategies usually  employed by administrators;

• Full backup

• Incremental Backup

• Differential Backup

• Copy Backup (Preston, 2007 p. 63).

Full backups backup an entire dataset irrespective of whether it has been altered or not. When the backup completes the saving process, all archive bits are cleared. Many organizations undertake full backups on a daily basis. Saving the entire data set means that the backup times are longer although restorations are quicker because the dataset resides on one tape (in the case of tape systems). Differential or incremental backup may be used together with full backups to reduce the amount of data being saved and thus to reduce time used to backup data.

Incremental backups only back up data that has been changed or that still has an archive bit stored on the file. Incremental backups are performed on a daily basis, while full backups are performed once per week. Therefore, incremental backups have different datasets that are stored and all the tapes used for backups are used for data restorations. After a full backup, the archive bits are cleared on files.

Differential backups can be used together with full backups just like in case with incremental backups.  “The difference is that this type of backup does not clear the archive bit which means that it continues to back up the changed files even after a full backup has been completed” (Turban & Volonino, 2011 p. 16).

Copy backups are full backups without a reset of the archive bit. They may be used prior to an operating system patch installation. In cases if the patch causes a system failure leading to data loss and requiring a reload, the dataset can be restored to its prior status without alteration of the archived bit.

Types of Backup Strategies

Some of the most common strategies for data backup include: Current work off-site backups that are done at regular intervals. Most of these backups are made on tapes. Another strategy is when backups are copied on disks and copies are saved to an off-site backup disk. The third strategy involves the Storage Area Networks (SAN) technology. Under this strategy, “data replication is done on an off-site location where in case of recovery, systems are the ones synchronized or restored but not the stored data” (Mohammad, et al. 2014 p. 75). Another strategy is the use of hybrid cloud solutions that replicate both off-site and on-site data centers.

According to this strategy, many organizations choose to use outsourced data recovery providers via cloud computing. This strategy is effective in that it allows continuous access to data and systems even after a disaster has occurred. 

Disaster Recovery

 According to Rajagopalan, Cully, Connor & Warfield (2012), the idea of implementing a pre-disaster plan in your organization, has a more cost effective outcome for the business in the long run as opposed to costs incurred for disaster response and recovery interventions. When an organization is implementing a disaster recovery plan, there are several issues that need to be considered. The main one is that historical data should be retained. This historical data is as current as the last database backup (p. 233 para 5)

Identifying Critical Data

Many organizations have proposed solutions that enable the automatic identification of critical data based on data dependency relationships and usage characteristics. Identification procedures of critical data are important in any organizations as the “modern distributed systems may contain a constantly shifting collection of data that may not be readily identifiable as critical by staff and non-experts” (Hutchinson & Springer, 2010 p. 96). Due to this, this paper examines the automatic identification of critical data using automatic identification systems. Real programs are used which develop inference algorithms that take un-annotated programs or partially-annotated items as input and locate critical data that need to be protected. This technique combines dynamic and static inference methods to minimize the likelihood of false identification of non-critical data as critical one.

Conclusion

The backup strategies discussed in this paper have been indispensable for many years. Notwithstanding, variations of these strategies have recently emerged. For example, system imaging is used to replace dataset backups. Backup systems are still used in many organizations to undertake complex tasks. It is recommended to use backup systems which are simple to understand and operate. This is achievable if the tasks are definite and well documented and the management of the system is done properly. In case of disaster recovery, a disaster recovery planner firstly looks at an organization`s recovery time objective (RTO) and recovery point objective (RPO) for various business procedures. This is because disaster recovery is a subset of business continuity planning of an organization. Taking in consideration these objectives, a disaster recovery planner is able to map the organization’s business process with the underlying IT systems so as to come up with a comprehensive disaster recovery plan.