In-Depth

What Does Storage Want to be When it Grows Up?

A Computer Associates vice president for marketing of BrightStor products suggests storage management today is a child still in the throes of personality development.

• By Jon William Toigo
• 09/23/2004

Lately, I have been attending a few “wandering mini-conferences”—events that some larger storage vendors such as Computer Associates, Network Appliance, Quantum, StorageTek and others seem to be preferring these days to those big, centralized, trade-show-with-education-track conferences of yesteryear.

Participating in these events as a “vendor agnostic” featured speaker provides a great way to compare notes with attending storage managers all over the country—and to rack up a lot of air miles and hotel credits that come in handy when take my eldest son, Max, around the world to look over colleges and universities.

At 16, Max is already doing deep blue computer science and coding in several languages, primarily on Windows and Linux platforms. The public high school that he elected to attend a couple of years ago actually specializes in computers and robotics. I confess to having been a bit nervous when he chose the school for fear that it would limit his exposure to other areas of study. Seriously: how many kids can you think of who actually know what they want to be when they grow up—especially at the ripe old age of 13 or 14?

I find myself confronting the same anxiety and asking a similar question when it comes to storage management technology: what exactly does storage management want to be when it grows up? The question surfaced recently as I listened to a presentation by Bob Davis, a Vice President at Computer Associates who oversees field marketing of the company’s BrightStor storage management product family. Davis makes it sound like storage management is a child that is still in the throes of personality development.

If I had to create a PowerPoint slide to illustrate his view, it would likely take the form of a human growth chart. At the far left would be a picture of a diaper-clad baby just beginning to crawl.

Think of the image as a metaphor for storage today. Beneath the picture would be the words: “Provision storage manually. Back up everything. No data classification or segregation.”

This describes the storage environment one finds in many smaller organizations. To the extent that storage is managed at all, it is mostly via the controls or utilities provided by the vendors on their own gear. The strategy works pretty well as long as there are only a few storage devices and only a few vendors. Baby-stepping consumers think that this is all the management required and believe that management is provided for free on every storage device they purchase. For many, the latter view is one that they never quite shake.

Next to the image of the infant is a picture of a preschool-age child, now out of diapers, and playing with blocks or with a kid’s computer game. Beneath the picture, add the caption: “For simplicity, buy one vendor’s gear and use the on-board management tools. Segregate the data from different applications into different partitions, zones, or onto multiple copies of the same device. Backup or replicate each data store based on a gut feel about application criticality and restoration priority in the wake of a disaster, or in response to poorly understood regulatory compliance requirements.”

This pre-adolescent strategy provides a sense of security and simplicity, especially as volume of data to be stored begins to grow. Not quite ready to go it alone, the adolescent is content to trade off flexibility for a warranty of care and service, even if it involves voluntarily locking oneself into a particular vendor’s gear—a strategy that may not make a lot of sense over time.

Now, we come to the third image: a picture of a person in his late teens, wearing a T-shirt sporting some idealistic cause or defiant gesture. I recently bought Max a T-shirt that reads: “Dad, it’s like having your own private ATM.” This is essentially the storage management philosophy of companies in this stage of development. These firms don’t like big storage vendors (“the Man”) and prefer to experiment around the edges, cobbling together lots of different gear from off-brand, no-name, or up-and-coming sources, regardless of its centralized manageability.

The upside of this strategy is that the storage equipment is usually less expensive to purchase and can be purpose-built to meet the needs of specific applications. Knowledge grows from experimentation with different interconnects, drive types, controllers, and so on, and a willingness to venture into some less-traveled paths, even if they ultimately prove to be dead ends.

The downside is that management costs mount over time, or interest in experimentation with storage shifts to other things, such as wireless or security or electronic music. Bottom line: all the promiscuous purchasing results in an unmanageable mess of infrastructure hosting data that is rarely backed up and poorly allocated across devices.

The end of the chart features two pictures of adults: one is a business-suited climber who has emerged from adolescence with a conservative—if reactionary—bent. He opts to replace his heterogeneous infrastructure with a single vendor’s SAN. He is less interested in storage technology than he is in “the big picture.” Storage is not a front-burner issue—business is. Opting for the path of least resistance, he buys whatever the “Cadillac equipment” vendors offer. Sure, the cost of the acquisition accelerates in a hockey stick curve after a year or two, but by then he may be promoted out of IT or working for another company altogether. He likes the fact that management is “free”—purchased with the hardware. Homogeneity is a no-brainer.

The other adult is a parent, possessed of that 10,000-yard stare reflecting wisdom gained through struggle and vigilance. To Davis, this is a symbol of what storage management needs to become in the future.

The parental manager takes the best from the heterogeneous, purpose-built infrastructure—which provides good data segregation and better storage-utilization efficiency than is possible with the monolithic offerings of three-letter hardware companies—and layers on an overarching, platform-agnostic, management-software framework to try to make all of the diverse components of the storage family work and play well together. Management consolidation drives labor cost out of the equation. Backups get done. The stage is set for consistent storage service together with controlled experimentation and improvement.

Being a family guy, I like Davis’ view a lot. Against the almost deafening noise generated by the industry around storage grids, utilities, ecosystems, federations, and the like, what we are really dealing with here is a bunch of kids who need some effective parenting.

What do you think? jtoigo@intnet.net