Tom Petrocelli's take on technology. Tom is the author of the book "Data Protection and Information Lifecycle Management" and a natural technology curmudgeon. This blog represents only my own views and not those of my employer, Enterprise Strategy Group. Frankly, mine are more amusing.

Friday, May 26, 2006

Just-in-time Data Protection

There is more activity in the data protection arena now than I've seen in years. We are reaping the fruits of years of new innovation to finally begin to achieve something special in data protection. Looking back over the past few years we have seen the emergence of:

  • Disk-based backup including VTL
  • Data reduction techniques
  • Continuous Data Protection (CDP) and real-time replication
  • WAFS or Wide Area File Systems
  • Tiered Storage
  • Data Lifecycle Management

All of these combined provide a whole host of choices in how one architectures data protection systems. It would seem that soon we will have the ability to protection all of our data all the time. But does that make the most sense?

Constant and complete data protection will always be very expensive. Bandwidth costs money, disk capacity costs money, and software costs money. To strive to backup everything all the time is to strive to waste money. Why? Because there is rarely a good reason to backup all the data, all the time? We may want to backup all of it sometimes. That the realm of tradition backup. We may want to backup some of it all the time. That's were CDP comes in.

Where we send it to be protected is also part of the equation. Do we keep it on-site? Copy it to tape to be stored off-site? Send it over a WAn link to another locations. All of these location options provide different levels of protecton at different costs.

What we really want is "just-in-time" data protection. Akin to JIT manufacturing, the goal is to protection data when it needs to be protected. Some data will need constant backup, perhaps because it is mission critical and needs to be available 24x7x365. Other data will only be backed up occasionally because it doesn't change much. Image files are an example of something that rarely if ever changes and only needs to be backed of once in awhile. Some data will be copied off-site to a DR facility immediately because it must be made available the second a disaster occurs. Other data will be archived and left in a vault on site. It all depends on how important the data is, how fast it needs to be accessed after a disaster or failure, and how much you are willing to pay for the priviledge of having the other two.

New data protection technology allows us to fine tune our systems to deliver data to backup exactly when it is needed. If we follow a philosophy of JIT data protection then we will soon find an equilibrium between protection, cost, and time. And that is a perfect combination. Start with the JIT goal and you have a fighting chance of achieving that balance. Do too much or pay too much and you will waste resources. Do too little or pay too little and your important data is always in danger.

No comments: