The State of Continuous Data Protection

Continuous data protection is one of the latest backup tools, yet few understand how it fits in with other products ... until they read this guide!

At one time, data protection and backup options were pretty sparse, consisting mainly of tape backup. With the advent of low-cost disk, storage managers now have an increasing array of options for protecting an organisation's data. Continuous data protection (CDP) represents one of the latest.

CDP allows data to be restored to the last data written before a system failure or to any point prior to that. Today, CDP capabilities are being included with snapshots and other forms of data replication and mirroring as part of data protection strategies that enable companies to achieve different recovery point objectives (RPOs), the amount of data the organization is willing to lose following a system failure, and different recovery time objectives (RTOs), the amount of time it takes to recover data following a failure.

CDP's beginnings

When CDP first appeared, it attracted a slew of startup vendors like Kashya (acquired by EMC), Revivio (now part of Symantec) and Topio (now owned by NetApp). They offered standalone appliances that provided CDP and other storage protection capabilities and were hailed for their ability to deliver an extremely short RPO at a much lower cost than the alternatives. As noted, many of those companies have been acquired by larger vendors that incorporated the CDP capabilities into their broader data protection offerings.

"CDP was aimed at a very small percentage of the market, at applications that required an RPO of zero. But how many applications in any data center really need an RPO of zero?" asks W. Curtis Preston, vice president of data protection at GlassHouse Technologies.

The technology as it was first envisioned also had other drawbacks. "Standalone CDP appliances added complexity and extra cost. It meant putting more pieces to be managed into the data center," says Greg Schulz, founder and senior analyst at StorageIO. "As a separate technology, it turned out to be dead on arrival for most companies."

CDP and data backup

It wasn't near-zero RPO but effective data backup that led the City of Safford to Asempra Technologies, a CDP vendor. It needed reliable data backup. "We had a backup server that went out to tape, but it just wasn't working right," says Derek Kruger, the city's IT and communications supervisor. Backup failures had already cost Safford considerable money. Frustrated, Kruger's team tried making drive images, but began running out of disk space. They brought in external hard drives, started rotating backups -- nothing they tried was effective.

Kruger started looking for something that would free his small IT group from the problems of tape backup and meet Safford's four-hour RPO. That pointed to a disk-based solution. Kruger looked into backup offerings from Asempra (Business Continuity Server [BCS]), EMC (RecoverPoint) and Symantec Corp. (Veritas NetBackup). The EMC product was overkill and beyond Safford's budget. Symantec would have worked, and at a cost of $US20,000, it even saved some money. However, he selected Asempra's BCS based on its ease of use, and the cost, $US30,000, fit within the budget.

With Asempra protecting 2 TB of Safford data, Kruger eliminated the tape system. Now the city is planning to install two new SANs and expects to use BCS to protect them as well. Compared to the costly failures of the tape system, says Kruger, "BCS paid for itself in the first six months."

CDP today

As the Safford experience suggests, CDP has become another form of data backup, often augmenting the existing backup process. "CDP is effective as a function, but it is not likely to replace other backup tools. It is really best as a feature in other backup products," says Lauren Whitehouse, analyst at the Enterprise Strategy Group.

After Symantec acquired Revivio, it incorporated it into its NetBackup product acquired with Veritas. NetBackup is aimed at large, multiplatform enterprises. Symantec also offers Backup Exec, a backup product for smaller Windows IT shops. "For Backup Exec we developed our own CDP capability," says Jason Fisher, director of product marketing for Symantec's Backup Exec group.

"CDP capabilities are overkill for 75% of our customers," says Fisher. Those who do use the CDP capabilities usually are protecting Microsoft Exchange and SQL Server applications or remote offices. With Backup Exec, the CDP capabilities are included with the core product, at no extra charge. Agents, however, must be deployed on the servers whose data is being protected.

EMC's RecoverPoint provides CDP based on technology it acquired with Kashya. It also offers a set of data protection tools, including array-based snapshots and replication. Also, their acquisition of Avamar brought deduplication technology. EMC's target for RecoverPoint is large enterprises.

Although CDP may not be thriving as a standalone product, it continues to attract considerable market attention. IBM announced an agreement in April to acquire FilesX, one of the few remaining standalone CDP players, and incorporate it into its Tivoli software. Similarly, Double-Take Software, which acquired TimeSpring and its CDP product in December 2007, relaunched it as TimeData, a CDP tool for the Microsoft Exchange, SQL Server and Windows environments.

Asempra and InMage Systems, at the time of this writing, remain independent CDP players, although Asempra regards CDP as just one part of its larger data protection offering. Going forward, that's what CDP likely will be for almost everybody, vendors and users alike, something to augment a broader data protection strategy.

Read more on Data protection, backup and archiving