A new approach to protecting data: Time Addressable Storage - Storage Networking

Computer Technology Review, Dec, 2003 by Kirby Wadsworth

Aren't Incremental Snapshots Good Enough?

It's clear that full mirror-splits are dead--way too expensive, clumsy, and limited in recovery point objective. But what about that new generation of incremental snapshot systems? They offer reduced capacity, and claim to only store the modified blocks--sounds like TAS, right? Not even close. Snapshots require pre-planning (you establish a schedule) so the recovery point is not what you want it to be, it's just the nearest available one. And, you still have to manage all the SPIT copies (can you say, headache with a capital H?). Snapshot systems typically suffer from significant performance degradation as the number and size of snaps increases. Snapshot copies are typically read only--meaning you have to copy them to live disk to effect a recovery--so they don't actually save space, and they definitely don't save time. TAS offers instant access to fully read/writable datasets--no moving, no copying.

[FIGURE 3 OMITTED]

Think of it this way--trains were okay until trucks came along. Snapshots are like trains--they take your goods to stations they choose, on their schedule. Trucks take your goods exactly where you want them delivered, exactly when you want them to get there. So, perhaps the new generation of more efficient and more powerful snapshot systems is like diesel trains--a lot better than the steam engines they replace, but still no match for the internal combustion engine.

It's simple--TAS offers better protection at lower cost. TAS systems should be included in ILM strategies as a perfect first line of defense against data loss and data corruption--freeing resources and offering compelling value. You owe it to yourself to find out more.

Kirby Wadsworth is vice president of Marketing and Business Development at Revivio (Lexington, MA)

www.revivio.com