Back Up Your Work, One Keystroke At a Time

Conventional data-backup practices have one gaping problem--they don't capture changes made to data between backups. A new approach called "continuous data protection" is designed to instantly copy and back up all changes to data, although it probably isn't needed by most companies.

The idea is to record every change to every data element--be it a Word document, a database entry, or an E-mail--thereby eliminating the risk of lost data. The technology is offered by large systems vendors such as IBM and EMC, as well as niche providers such as Asempra Technologies, Revivio, and TimeSpring Software. It will appeal to businesses such as financial-services and health-care companies under pressure to guarantee against lost data, while also maintaining a high degree of application availability. For most companies, however, daily or hourly backups are sufficient.

EMC last week unveiled RecoverPoint, software that captures all changes as they occur and copies them to a recovery storage space. It's platform-independent and lets groups of related applications be restarted from the same point in time, says George Symons, chief technology officer at EMC's information-management software group. Prices start at $75,000. FalconStor Software Inc. also expanded its continuous-data-protection services last week to include host machines' direct-attached storage via its DiskSafe replication software. DiskSafe transmits and records changed data blocks to meet specific recovery objectives.

Magellan Health Services Inc. is implementing disk-to-disk backups using EMC's TimeFinder snapshot software in its Windows environment, and may extend it to Unix. It's also evaluating EMC's RecoverPoint and will probably deploy it, says Bob Odenheimer, Magellan's senior VP of IT operations.

Because continuous data protection only records changes to data, there's less data to be copied. That shortens the time it takes to recover from a failure, says Brian Babineau, an analyst at Enterprise Strategy Group, and reduces the amount of data lost because of a failure.