This article can also be found in the Premium Editorial Download "Storage magazine: What you need to know about all solid-state arrays."
Download it now to read this article plus other related content.
I’ve been ruminating on this bit of history for several reasons. First, I had to endure EMC World and all their chatter about Atmos and clouds. Then, I had to deal with the planners of a prominent fall trade show regarding an email blast they were issuing to promote the show that was literally brimming with marketing bull that hit all the shiny new memes: cloud, mobility, BYOD (bring your own device -- I had to look it up), “big data” and so on. And lastly, I found myself ticked off by an email I received stating that data is doubling every 1.2 years to justify the purchase of faster servers.
Let’s go to the last one first. A prominent server manufacturer, using money provided by a prominent manufacturer of the latest generation of fast CPUs, asserted a rate of data growth that was attributed to a smart guy at MIT Sloan School of Management. I copied and pasted the link and went to a blog page where the factoid was gathered and found nothing but a statement made by the bearded professorial writer -- an assertion of this data growth stat. There was nothing to justify it; no empirical evidence whatsoever, just a statement I guess we were supposed to take as gospel because it was wrapped in a nicely designed blog site at MIT Sloan.
I’ve tracked data growth claims since IDC began writing “Exploding Digital Universe” research papers with the expectation, I guess, that they would convince consumers to buy more capacity for their SAN-that-isn’t-really-a-network.
I know of no empirical research that demonstrates with any validity the rate of data growth in a business. That’s because no one knows how fast their data is growing; they only know how much storage they’re buying year over year. The latter isn’t an indication of data growth rates, but of data mismanagement, pure and simple. A combination of archive and good old-fashioned data hygiene could whittle that capacity demand down substantially, based on the research of my firm, which demonstrated in a population of more than 3,000 companies that less than 30% of the data occupying disk needs to be on spinning rust.
But no one wants to talk about a storage reclamation strategy, only a storage expansion strategy. Everything-to-disk (soon to be everything-on-flash solid-state drives) ideology has replaced common sense thinking about the data deluge and I’m frankly getting sick of it.
Why do you need storage clouds like Atmos? As elbow room for your SANs and network-attached storage (NAS) boxes, of course. And why do we need the elbow room? Because of all that big data stuff we should all be doing to make our businesses more competitive, or to redress the piggishness of server virtualization wares that require data to be replicated all over the place to support high-availability clustering -- not only for critical apps, but for file servers and low traffic Web servers.
Omnia in orbis -- Latin for “everything on disk” -- may sound collegiate, but it isn’t a foundation for sound storage strategy. It isn’t the basis for buying a high-end server. And it shouldn’t be the organizing principle for a trade show that once served as part primer/part technology update for practitioners of the dark arts of IT.
BIO: Jon William Toigo is a 30-year IT veteran, CEO and managing principal of Toigo Partners International, and chairman of the Data Management Institute.
This was first published in August 2012