How to Prevent Data Loss Using Continuous Data Protection

Enterprises today are not limited to tape-based backups as their only means of data protection. A combination of tape backup, high availability and continuous data protection technology can now be used together to provide seamless protection against data loss, data corruption and system downtime. Knowledge Center contributor Doug Piper explains how tape backup, high availability and continuous data protection technology can provide the most complete level of data protection for your enterprise.


Enterprises have used tape as a backup medium for as long as they have been using computers. Apart from storing hard-copy source documents, tape backup was initially the only data protection mechanism available. Many organizations still rely exclusively on paper and tape backups today but that needs to change.

First, paper documents are no longer a recovery option because many transactions are fully electronic. Second, as 24x7 operations become more prevalent, and as regulators and other stakeholders require more stringent safeguarding of critical data, relying solely on tape is no longer sufficient. Finally, paper and tape provide inadequate recovery options in an age when systems connect to the world and are highly vulnerable to attack. Today, seamless data protection and instant recovery are absolute necessities.

Despite great increases in tape speeds, tape drives are still slow compared to disks. Recovering data and applications after a disaster can take several hours or even days, particularly if the tapes must be retrieved from an offsite vault. In addition, although tape recovery is much more automated than in the past, it still requires some manual processing that can result in errors and delays. Moreover, long recovery times mean that tape backups cannot serve to keep operations running during brief maintenance shutdowns.

More limitations of tape backup

Another problem with tape backups is that they allow the recovery of data to only one point in time, usually sometime during the previous night. Updates applied later are lost if a disaster destroys the data center. Worse, if last night's tape has not yet been shipped offsite, it may be destroyed as well, forcing the company to revert to an even older backup version. The same will be true if the most recent tape is unreadable. In addition, if a virus attacked the systems at, say, 4:03p.m., one might want to recover data to its state at 4:02p.m. But since tapes allow restoration of data only to the point in time when the backup was created, they do not provide that option.

At one time, operators had to stop production applications while running a backup job. Today, backup-while-active features theoretically allow online applications to function during backup operations but that is often not practical. Transactions caught in mid-flight during a backup operation may corrupt the data on the tape. Thus, half of a transaction may be saved but a balancing debit or credit may be omitted. Furthermore, because backup jobs grab every byte of data on the disks as quickly as possible, online applications may slow almost to a stop when backups are running. This wasn't an issue when backups could run during "off hours."

Today, when e-commerce and global supply chains require around-the-clock availability, running tape backup jobs on a production server is unacceptable at any time.