Tiering, ILM, HSM and WGAF

In 2004, SNIA defined Information Lifecycle Management (ILM) as comprising the policies, processes, practices, and tools used to align the business value of information with the most appropriate and cost effective IT infrastructure (I would have added 'for its placement' but the ILMers at SNIA are wicked fussy about inferring that ILM has ANYTHING to do with storage. –Ed.) from the time information is conceived through its final disposition. Information is aligned with business processes through management policies and service levels associated with applications, metadata, information, and data.

 I like that definition – in fact, I was hanging around SNIA a lot back then kick starting the CDP working group, and may have contributed to that definition. Who knows, sounds like something I might have written- lots of warm air whooshing around those words…
In January of 2005, an unattributed author at TechTarget opined that: Tiered storage is the assignment of different categories of data to different types of storage media in order to reduce total storage cost. Categories may be based on levels of protection needed, performance requirements, frequency of use, and other considerations.
In early 2006, SNIA added that "ILM is more than Tiered Storage" and went on to suggest the need for a complex set of data classification capabilities. Whoops – too little too soon? All the data classification start-ups I knew blew up or died on the vine.
Then, according to someone unnamed author at TechTarget back in March, 2006 the definition of Information life cycle management (ILM) became:
"A comprehensive approach to managing the flow of an information system's data and associated metadata from creation and initial storage to the time when it becomes obsolete and is deleted."
I'm not as crazy about that definition – final disposition not always being obsolescence and deletion and all...
The same unknown author went on to add:
"Unlike earlier approaches to data storage management, ILM involves all aspects of dealing with data, starting with user practices, rather than just automating storage procedures, as for example, hierarchical storage management (HSM) does. Also in contrast to older systems, ILM enables more complex criteria for storage management than data age and frequency of access."
I guess I don't really understand what 'all aspects of dealing with data starting with user practices' means. Do you? I do appreciate that ILM can use more complex criteria than age and access frequency, but I didn't realize that HSM couldn't…
According to the same source, this time written in July 2001 by Gaston Navea, the definition of Hierarchical Storage Management (HSM) is:
Policy-based management of file backup and archiving in a way that uses storage devices economically and without the user needing to be aware of when files are being retrieved from backup storage media.
Gaston then adds that policies might include age or access, but also states that executables might be excluded, inferring at least that 'type' metadata is also a valid differentiator. But for me, Gaston puts the stake in the heart of HSM when he goes on to say, "The apparently available files are known as stubs and point to the real location of the file in backup storage." No wonder HSM flopped. Stubs are a nightmare. Disconnecting the location from the payload is a huge strategic mistake. Either is useless without the other, the opportunities for disconnections are ample, and the cost of disconnect is horrific. So HSM is out.
Back to the ILMers at SNIA. They believe ILM represents a holistic approach to information management. I suppose you have to appreciate anything that represents a holistic approach. But then they go on to introduce ILM 2.0 which re-introduces the concept of Information Lifecycle Management with accompanying processes and procedures absent from earlier days. I personally hate all 2.0 terminology, but that's just me.
And they warn us, even with ILM 2.0, to: Make no mistake; Information Lifecycle Management is still a difficult and challenging proposition. They go on to define ILM 2.0 as: a service management style framework for cost-effectively aligning datacenter storage, security, services, applications, and infrastructure with the business requirements for the organization's information. This model, by the way, as 100 individual elements, and comes with a checklist of 21 specific steps to follow to achieve it.
As Andy Azula says in the recent UPS whiteboard commercial, "Is anyone else getting thirsty?"
Referring to an article by George Crump on file virtualization, my F5 colleague, Don MacVittie wrote, "The other thing that made me a throw up a little in the back of my throat was his use of the dread phrase "ILM" (Information Lifecycle Management). I shudder when our marketing organization uses it too." This is memorable, and pretty witty if you ask me, especially considering that Don is officially in marketing at F5… That snark aside, Don's core point is that many of the better elements of the ILM concept have survived the gooey-glop-ulization of the term itself in the form of real solutions to actual customer problems. This takes us back to Tiering.
Tiering works. Plain and simple. If you don't believe me, watch this short powerful video of one of our customers, RHWL Architects, discussing what tiering did for them. He says something like, "All our storage problems disappeared overnight". Pretty powerful and not goo-gloppy at all. I happen to know that these folks did not follow any 21 step, 100 element service management style framework to solve 'all their storage problems overnight'. They implemented an ARX file virtualization appliance and tiered their storage – that's all.
How could anything be that simple? Ah, heck, we storage marketers have been complexifying and goo-glopping this whole issue for a decade or more. Somewhere along the line, somebody got the idea that if we make it hard, complex, and scary, then we can charge more money to fix it. (maybe this is all really my fault afterall.)
Look it's really not complicated. Sort your files. Then put some of them on one array and some on another. The rest starts to take care of itself after that. Sort them anyway you like, by age, size, owner, extension, type, whatever. Most people use 'last modified date' first, and then get more granular later, but you will figure that out on your own as you get smarter about tiering.
Tiering is sort of akin to contributing to your 401K (in USA that's a retirement savings account). You don't need to think about it too hard. Just do it. Not doing it is really stupid, and anyone who tells you not to is an idiot or worse. You don't need to have a perfectly balanced investment strategy, and a complete understanding of your financial risk profile, before you sign up for payroll deductions. All you need to know is it saves you money on taxes and gets you started on a nest egg. Once you start building up experience (and capital) you can add more finesse. Same here.
Just do it. Sort, Tier, Save. Simple – does not require arguments about definitions, and will not make you throw up in the back of your throat.
PS – Make sure you use a solid solution - like the F5 ARX - to do it. Don't use stubs. Don't use half-baked hybrid clunkers that go in and out of band (and rely on stubs even if they say they don't). Investing in a 401K works, letting Bernie Madoff manage it doesn't. Be smart. Do the right thing.
PPS - I am leaving it to you to figure out the WGAF acronym. Use your imagination.
PPPS - Admit it, you didn't know who Andy Azula is did you? See, you always learn something reading this blog...maybe not about storage, but something...