How continuous data protection works

* Testing out IBM's Tivoli Continuous Data Protection for Files

Time-based backup and recovery - full backups, incrementals, and their various permutations - have been with us forever. Now a new approach, continuous data protection, looks to be much better and is rapidly changing the backup and recovery landscape.

CDP captures every change made in a file, keeping a record of the changes. The record is available to users, who may then review each change event and initiate a recovery from whichever point they wish. This illustrates the fundamental difference between CDP and traditional backups: CDP is event-based.

Certainly, there are many ways for vendors to deliver CDP. For the last month, I've been looking at a CDP product from IBM called Tivoli Continuous Data Protection for Files. It's simple, it's cheap ($35 per client license), and it works. Here's how it works.

The first step is to specify all files to be protected with the CDP software. During step two, you identify and allocate space on the target device, where the backups will be stored. The target may be local or remote. In either case, the more space you set aside on the target device, the farther back in time you can go to restore your files. You can also specify files to be excluded from CDP protection (system files and other backup files are likely candidates for exclusion). These can be protected by a scheduled (and thus non-CDP) backup, or left unprotected.

If you go with the remote option, you have a choice between backing up to a remote disk, to a WEBDAV (Web-based distributed authoring and versioning) device, to a remote disk, or to IBM's TSM backup software, located somewhere on a corporate server. This set of choices here means that the product should work in everything from enterprise to small office situations. How appropriate it is for your needs? You'll have to decide for yourself.

My first tests were to an external USB drive; I had planned to test the remote capabilities later, but the release notes indicate that an external USB drive is, apparently, considered "remote" despite what Windows has to say. That's a bit quirky to be sure, but the system seemed to work just fine when I defined the external drive as local.

When first launched, the product does a full backup, a process no less time-consuming in a CDP environment than in any other. Once that is finished however, the benefits accrue.

A CDP backup takes place in the background, with negligible impact on either CPU or memory usage; the only time you have to pay attention to the program is when you need to do a recovery.

When that happens, fire up the Java-based interface, go to "Restore", and locate the file you just blew away. You will likely see several versions listed on the restore page, one for each time you did a save. Most of the time you'll want the most recent version, but if you are not sure, click on any/all the ones shown and decide which works best for you. It can then be saved back to its original directory with a unique file name.

The user interface is just fine, although my version ( certainly would have benefited from a thorough good going over by a technical editor. The online documentation is scanty, but sufficient. I submit that it is a good thing when scanty documentation suffices.

Tivoli CDP works for Windows file servers, desktops and laptops.

Enterprise Management Associates is researching various vendors' CDP offerings. I'll be looking at many other CDP products over the next several months as I head up a project. Let me know if you'd like to be kept informed.

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.
Must read: 10 new UI features coming to Windows 10