Over the last year more than a dozen start-ups have climbed into the so-called continuous data protection (CDP) market, and the last few months saw the entry of heavyweights such as Microsoft, IBM, Symantec and, most recently, EMC (see “EMC ‘validates’ CDP,” on the cover).
However, putting aside the argument of what’s “true” CDP and what’s “near” CDP, end users don’t appear to be flocking to this technology, despite the market validation by the big guns. A number of surveys indicate lukewarm interest at best.
For example, in a survey of more than 200 InfoStor readers (conducted by Toigo Partners’ Data Management Institute and commissioned by Topio) one-fourth of the readers didn’t even know what their company’s position was on CDP (see figure). And all of those survey respondents were storage professionals!
In the preliminary results from a separate InfoStor QuickVote reader poll, respondents were about evenly split among “have already implemented,” “will implement in 2006,” “no plans for CDP,” and “do not know what CDP is.”
Bingo. I think part of the problem here is the term itself and lack of consensus on what it means. Basically, CDP backs up every time a change is made to the data and users have the ability to recover-rapidly-to any point in time (as opposed to, say, daily or weekly backups or recovery based on periodic snapshots). Despite the best efforts of outfits like InfoStor and the Storage Networking Industry Association (which has a group dedicated to CDP), I think a lot of IT professionals are still a bit hazy on this technology.
For example, I’d bet that for many small to medium-sized businesses (SMBs) weekly backups with daily incrementals constitute “continuous” data protection, because until CDP came along that was the time-honored standard for diligent backup practices.
Further confusing the issue: Some CDP products are file-based and some are block-based, and there’s a huge gap in pricing between “near” CDP software such as Microsoft’s Data Protection Manager (DPM), which costs less than $1,000, and “true” CDP software/appliances that would bust the budgets of most SMBs.
It’s likely that in terms of end-user adoption the CDP market will fracture along the “near” versus “true” CDP lines, which translates into SMB (small budget, moderate recovery requirements) vs. large enterprise (big budgets, stringent recovery time/point objectives). My guess is that the under-$1,000 CDP products will be the hottest sellers because hourly snapshots and restores point are “good enough.”