Secure Data with Simple Disk Mirroring

I'll bet you have at least one or two SMB clients who constantly worry about losing their data. It's natural for them to worry, because data is the backbone of many businesses. Fortunately, today's high-capacity, network-accessible add-on drives for centralized storage with Ethernet connectivity make automating backups is an effective solution and a lucrative add-on. Even better, they're a breeze to set up.

In this Recipe, I'll show you how to add an extra layer of protection to your client's data using RAID 1 (also known as disk mirroring) on a slick new network add-on box from Maxtor that offers dual 500-MB hard drives, and a fast 1-Gigabit Ethernet network interface that's ideal for the job.

As noted, we all have clients who wring their hands about potential data loss. Sales types, for example, may fret over the possibility of losing data on their largest accounts. Office managers and sys admins are concerned about losing company data. Truth is, they may indeed have a good reason to worry, even if they use automated backups.

As convenient and reliable as desktop or add-on drives are for backups, they are still of the same design as the mechanical devices you are trying to protect. All magnetic drives are subject to the laws of physics that lead to disk failures. That's the reality: Any magnetic drive, by its nature, is subject to failure. More precisely, every disk will fail at some point. The question is when.

For this reason, businesses with critical data should do more to protect their data. They may require RAID 1 protection, which essentially "mirrors" all their data to a second disk as a second layer of protection. By mirroring data, RAID 1 continues to make all data available immediately after a failure, so critical applications can continue to run. If you or your client's business can't interrupt business to restore missing data from backups, RAID 1 is the answer.

Until now, adding RAID 1 meant dedicating and configuring a PC with a drive array. But times are changing. RAID is now becoming available in convenient network add-on storage boxes.

Big Uses for Large Drives

Not too long ago, when hard-disk manufacturers unveiled the first 500-GB drives, I remember thinking, "What could be better than this?" But soon, manufacturers were installing these devices in desktop boxes for easy-to-add external storage. Equipped with 100 megabits-per-second (Mbps) Ethernet ports, these sleek boxes made gobs of Network Access Storage (NAS) capacity available and affordable for small-business networks, even standalone machines. In addition, easy-to-use software for automating backup was added that would let these central storage units act as automated backup boxes, thus doing away with tapes and other messy backup solutions.

Now I was thinking, "What could be better than one reliable 500-GB drive in a quiet, easy-to-install desktop box?" Simple. Two 500-GB hard drives in that same package! In fact, this past September Maxtor began delivering its Shared Storage II dual 500 GB device—yes, that equates to a 1-terabyte (TB) add-on—with fast, Gigabit Ethernet network connectivity. Shipping with the box is easy-to-install and configure software to configure the dual drives into an impressive 1-TB RAID 0 array or an ultra-safe RAID 1 mirrored configuration.

Two Disks are Better than One

Simply put, two disks are better than one. That's because from two physical 500-GB drives, you can build a RAID 1 array to have one disk mirror the other. With a RAID 1 array, you're not just backing up one point of failure to another point of failure, but to two physical drives. I'm not a statistician, but having data in three places—the workstation, backup drive, and mirrored image—makes that data almost impossible to lose. Plus, once in place, mirroring is essentially invisible to users. Should a disk fail, users see a pop-up message indicating that a disk has failed. But the mirrored data is still available and, most importantly, applications keep running.

First-time users of RAID 1 are typically surprised by how little work is needed for set-up. And once disk mirroring is set up, it requires no maintenance. There's the initial configuration of the RAID 1 array when first installing the drives. But after that, mirroring requires no additional actions or maintenance to keep a constant vigil over your client's data.

Alternatives Can Be Messy

While RAID 1 implementations require more disk space than optical storage and other alternatives, they are easier to design and install. For example, high-capacity DVDs provide a reliable format on which to preserve important data, and they may be necessary if backups need to be stored off site. But going to a typical DVD backup regimen on a regular basis soon begins to look like the old tape-backup routine. Inexperienced backup administrators quickly learn that the constant shuffling, labeling and storing of DVDs is a pain.

If you're like me, in addition to CD/DVD strategies, you've worked out other methods to add an addition layer of protection to backups. These include rotating removable drives and deploying multiple external devices. But each drive needs to be carefully handled and marked to keep backups straight. Methods like these may be unavoidable if you must store data off-site. But if your data is staying on-site, it's hard to beat the convenience and ease of disk mirroring.

About RAID 0 and 1

RAID provides a way to logically connect two or more drives into a group, or "array," for better performance or better protection. There are several flavors of RAID protection, some for use with larger arrays with many disks. The RAID type best for your use depends on the type of application you are running and whether you need greater data security or performance.

RAID types 0 and 1 are the types most likely to be implemented on NAS add-on boxes for PCs and small networks. Here's a look at the benefits and drawbacks of each:

* RAID 0: Also called "disk striping." All data is spread out in chunks across all the disks in the RAID array. Performance is great because you spread out the load of storing and retrieving data onto more physical drives. There is no parity generated for RAID 0. Therefore, there is no overhead to write data to RAID 0 disks. But RAID 0 is best for better performance, not high availability, since if either disk fails, data is not available for applications. In essence, RAID 0 effectively combines the storage of the available drives and boasts the fastest performance.

* RAID 1: Also called "disk mirroring." All the data is written to at least two separate physical disks. Once RAID 1 is running, the disks are, in effect, mirror images of each other. Should one drive fail, the data remains available from the other drive. This approach provides the highest level of data protection, but requires the most disk space—exactly double. For example, to protect a 500-GB disk, you'll need another 500-GB disk in the array. That said, mirroring is ideal for fast read operations: Data read performance is almost doubled, as consecutive reads can occur on each drive at the same time. But writing to disks is slightly slower, since the data needs to be written into each disk.