Continuous data protection has changed the way modern organizations back up data. Even though the technology has been around for quite some time, it took a while for it to gain traction. Today, CDP has almost completely replaced traditional backups.
Even so, there are both advantages and disadvantages to using CDP. Backup admins must know if the organization's budget, operating system and storage method are compatible with CDP before fully committing.
For some organizations, continuous data protection is a no-brainer. But closing the backup window might not outweigh the high hardware costs or the risk of failure if organizations implement CDP incorrectly.
Closing the backup window
One of the major continuous data protection advantages is that it eliminates the backup window. Rather than perform a monolithic backup each night, as was once common practice, CDP backs up data on a nearly continuous basis throughout the day.
Not only does this eliminate the backup window, but it also dramatically shortens the recovery point objective (RPO). The RPO is related to the frequency with which backups are taken and essentially determines the amount of data that could potentially be lost in the event that a backup needs to be restored.
When a traditional backup runs once each night, any data that accumulates the next day -- before the next backup -- can potentially be lost if a failure occurs. For example, if a backup runs at 11 p.m. and a failure were to occur at 4 p.m. the next afternoon, then the organization could potentially lose up to 17 hours' worth of data.
With CDP, a failure should never result in the loss of more than a few minutes' worth of data because the backup runs on a nearly continuous basis. Although CDP backups also create recovery points, they do so far more frequently than would be the case for a nightly backup. Some modern CDP tools can create recovery points as frequently as every 30 seconds.
Compatibility and cloud issues
Organizations that are interested in CDP must consider application and OS compatibility. Like any backup technology, CDP can only work if it is compatible with the resources that need protection. Luckily, most modern CDP products are designed to work with all commonly used OSes, applications running on premises and even SaaS applications.
Organizations that use cloud storage might want to reconsider adopting continuous data protection. While it is theoretically possible to back up cloud storage using an on-premises CDP tool, issues such as latency and data egress fees can make such backups impractical.
Don't let hardware costs sneak up on you
Another issue to consider is cost. Continuous data protection tools are disk-based. As such, the organization will likely have to make a significant investment in physical disk storage before deploying CDP.
Keep in mind that continuous data protection requires more than just raw storage capacity. The organization's storage must be fast enough to keep pace with the data's change rate, even during periods of peak activity.
This generally means that backup data will need to be striped across many disks in order for the CDP tool to provide the required number of IOPS. Similarly, the organization must ensure that it has enough network bandwidth to run continuous data protection.
Finally, it is critical to make sure that a CDP server does not become a single point of failure. Organizations will need a secondary means of protecting the data they have backed up. Otherwise, the CDP server itself could become a single point of failure.
Many organizations replicate their CDP data to a secondary CDP server that can take over should the primary server fail. Another common approach involves using either a disk-to-disk-to-tape or a disk-to-disk-to-cloud architecture.
These architectures use two CDP servers with duplicate data copies for redundancy, and periodically also back up the data to either tape or cloud storage. These tape and cloud backup copies are not created in real time and might only occur once a day (tape) or a few times a day (cloud).