You run around with your head on fire long enough, and sooner or later you forget about the guy throwing matches at you and look for a hose. We are tactical devils, us storage people, looking to fix the result as opposed to correcting the cause. We worry about backing stuff up, but not recovering it. Worse, we worry about disks--and not data. The data is our raison d'etre. It's why we're employed, or at least should be.
I was at the Immigration and Naturalization Service a few weeks ago--we are in process of adopting a Chinese baby girl. They now take digital pictures of your fingerprints--each becomes a unique object, and is stored in the mega-super data bank somewhere down south. That got me to thinking: How do they get real-time info when that database must contain millions and millions of objects? I bet the government has much more smart people spending their time thinking about how to house the fingerprints than how to use the information contained in those fingerprints. Illogical, maybe, but it does sound like the norm.
We spent over $3 billion last year on backup technologies--but is the stuff we back up (over and over again, typically) valuable to our business anymore? We keep more data online than ever before--hooray! I guarantee you we keep more useless data online than ever before as well.
There are two key concepts to grasp here, people--the value of data, and where to put it. I used to tell people (correctly, mind you) to first figure out what is out there, and where. You still need to do that. Most of you don't really know. You have a good idea, but there will be surprises, I can guarantee it. If I really press you, you'll sound as specific as an Arnold Schwarzenegger campaign speech. Second, once you ascertain what is out there--and you've weeded out the live bootleg "My Sharona" MP3s--you need to take a first pass at a value pecking order. It will help you continue the weeding process.
Once you create a value hierarchy, you can then start to think about tactical things such as, is the data on the most appropriate asset or not. If you value your engineering work in process higher than your contact management system, then why is the code data sitting on JBODs scattered around the building while the corporate ACT database is sitting on four tons of heavy iron?
Information Lifecycle Management (ILM) is the combination of policy, process and technology to physically place data on devices based on objective and subjective measures. Objective is easy--how often is it accessed, how old is it, etc. Subjective is hard. Subjective is value. Just because we rarely access something, doesn't mean it isn't valuable and vice versa.
The automation of data movement/migration based on our objective and subjective policies and best practices is what ILM is supposed to do--but it never eliminates the need for human intelligence to create those polices and practices. The value that IT can bring to the organization is just that--the intellectual process of assigning value and creating the best practices to deal with different data types and values, at different points in time. All data is not created equally, and it certainly doesn't stay the same on the value/time continuum. Strategic thinking is what IT needs to elevate itself inside the organization, not more tactical disk garf.
- Taming Hadoop: Storage Tiering for Big Data –Western Digital
- The Emergence of a New, Highly Flash-Optimized Tier 1 Storage Platform –Pure Storage