When considering whether you need a full copy of your data elsewhere, you need to understand the data loss scenarios you're planning for, and whether they're logical or physical. Where CDP offers protection against physical loss or corruption, it further provides the ability to protect against logical corruption. Logical corruption is the bugbear of many data protection strategies which simply replicate data at the block level in whatever condition it exists. This means that if the primary data becomes logically corrupted then that corruption is replicated to your secondary target rendering it useless for recovery.
CDP offers similar benefits to snapshot technologies provided with array replication, but has greater granularity by providing point-in-time recovery based on the configured retention settings. Point-in-time recovery gives you the ability to restore data to a point prior to when corruption occurred. This provides a data protection strategy that not only protects against physical loss, but further protects against logical corruption by restoring the system up to the last write before that corruption occurred.
The type of data you're protecting will ultimately determine the continuous data protection strategy you require. A block-based CDP product will protect any application and associated data stored on that volume. An application-based CDP system is designed to protect only one application type, regardless of where it's stored. The former introduces greater versatility but typically does so at greater cost, while the latter tends to be less expensive with enhanced functionality designed for a specific application type.
CDP offers effective protection against logical and physical corruption, with a number of continuous data protection products offering true and near CDP in the same package. However, whether this is suitable as your only data copy, is ultimately up to your recovery time objective (RTO) and recovery point objective (RPO) metrics combined with compliance and regulatory governance that determine how and where you need to recover your data.
This was first published in September 2009