Chapter 11. Conclusion: Joining the Quest for the Holy Grail

As this conclusion is being written, a news article has just come to my desk from somewhere out in the World Wide Web. The headline announces the discovery of an alien solar system with a planetary structure similar to ours. The lede to the article asks the always-provocative question of whether there may be a world like ours, populated with people like us, orbiting that distant star.

My take: If there is another Earth out there, I hope it has a few data managers. If so, we might be able to make an exchange of some sort , obtaining in the process some best practices and standards-based solutions that we need so desperately here on our planet.

At the beginning of this book, I noted that data storage is the most expensive line item in most corporate IT budgets , and it continues to grow irrespective of the condition of the general economy. There are several contributing causes to this phenomenon , including the apparent propensity of storage vendors to release non-interoperable and half-baked technology to market in order to get the jump on competition, and the prevalence of poor storage management tools that provide insufficient capabilities to enable us to manage burgeoning data with greater efficiency.

Ultimately, however, it is the absence of real data managers that best explains the phenomenon. Consumers are really to blame.

Today, most organizations entrust the design, management, and administration of their most prized and irreplaceable asset, data, to personnel who have none of the necessary prerequisites to do the job. Storage management is typically an afterthought: a set of tasks lumped onto the shoulders of IT technicians and systems administrators who 1) don't particularly care about data storage or data management, and 2) generally possess no specific skills or knowledge for performing those tasks. This isn't always the case, but more often than not those who manage storage for a living have zero training to do their jobs.

Under the circumstances, it should come as no surprise that storage today is so expensive, or that our storage platforms are oversubscribed and underutilized , or that contemporary data management is such a mess.

To be honest, with only a few exceptions, the field of data management doesn't exactly attract the best and brightest minds. Architecturally, storage may have been one of the three fundamental components of the original design of the von Neumann machine, the modern computer ”the other two being CPU and memory. Historically, however, it was almost always been viewed as the weak sister of the three.

Slow and ugly and cumbersome, disk and tape storage has about as much "sex appeal " as a garbage truck. Historically, in mainframe culture, we put at least two layers of random access memory between the processor and the magnetic disk and spent far more time trying to figure out how to optimize these dynamic RAM caches than we ever did pondering the most efficient way to manage data dumped onto magnetic media platters. If DASD was the proverbial dump truck, central or expanded store RAM were Formula One racecars. Ask any computer engineering undergrad which one he or she would rather drive. You can probably guess the answer.

Just as storage technology is not the most sought-after degree program in technical colleges and universities, it is also a discipline with enormous responsibilities and few perks in the "real world" of work. In most contemporary IT organizations, if you are unfortunate enough to be tapped with storage management duties (doubtless because you were the slowest person to duck into your cubicle when they were handing out the assignment), you live a relatively thankless life. Users become aggravated with you whenever an application stops because of inadequate disk space, but relatively few dedicate any time to thinking about the consequences of all the email they refuse to purge or archive.

If you are tasked with storage planning, administration, and management, you probably have permanent footprints on your back from all the vendors of large-scale storage arrays or Fibre Channel SANs who, after you carefully and diligently reviewed their offerings and rejected them for your application requirements, decided to climb over your objections and make their case directly to the CFO. And, instead of being commended for looking out for the company's money and serving its needs for reliable storage, your efforts are rewarded by two guys who show up at your data center door with dollies asking where they should put several big storage arrays that you didn't order. The purchase order is signed by a senior business manager who, to paraphrase the old saw, doesn't know storage from Shinola.

In general, storage folk labor in isolation and obscurity. Until very recently, there have been no storage-specific user groups, and mentioning that you managed storage at a meeting of a computer user group was likely to get you a blank stare.

"You're a database administrator?" asks one polite attendee , probably someone's spouse. "No," you reply, "I manage storage. You know: tape, disk drives , like that." You probably aren't very popular on date night.

To paraphrase Rodney Dangerfield, storage management earns you little respect. There are no formal job titles in most organizations for what you do. You don't have any professional identity within the IT field. There is no clear career path for you except that you must eventually subject yourself to human cloning in order to manage more and more storage capacity.

Your isolation works to the advantage of the vendors. When something doesn't work, they can tell you with impunity that the fault was all yours. In the absence of a group of peers with whom to compare notes, you have no way to question the validity of their claims.

It must make you angry when you hear a vendor at a storage conference refer to the many focus groups and customer surveys they have conducted that led them to add this or that feature to their latest product ”features for which you have no use. It must make you doubly angry when a vendor representative in a panel discussion says that the reason that a certain capability is not part of their product is because no one ever asked for it ”but you have been telling your account representative that you needed the capability in question for no less than 10 years .

If things are as bad as I am suggesting, some readers may ask, why are there storage folk out in the trenches who are actually wandering around with big smiles on their faces? Don't their backups fail? Aren't they also being overcharged for disk arrays that they are basically just a few shelves of commodity disk drives inside a tin shell? Haven't they been sold brain-dead fabrics that didn't do any of the things promised by the SAN visionaries? Aren't they also confronting 40 percent allocation efficiency (and substantially less utilization efficiency) even on a good day?

Sure they are. Though some are in what the psychologists call a "state of denial." John Tyrrell, one of the patent holders on Systems Managed Storage, once remarked that he participated in the first 100 studies around SMS done with customers. ^[1]

"When we added up the people time to 'manage' storage, we not only counted the IT people but we interviewed application groups and asked them how they did their job with respect to storage. They said, 'We don't do storage management.' But after interviewing them, we found they spent lots of time cleaning up space, correcting job failures, resubmitting, looking for space on volumes , changing JCL and resubmitting, backing up their own data, and so on. When we added up the real cost of managing, we found that 80 percent of the management cost was in the application areas, not the data center. The real cost of managing lots of islands of storage is astronomical. We finally stopped keeping track of it. If you think about the time you spend managing your own little laptop island, this will make sense to you."

Distributed computing has made everyone ”and no one ”a data manager. No one much wants the job, which involves in roughly equal parts ,

• Doing the manual maintenance chores that no one else wants to do with tools that trace their origins to some pre-computer era;

• Coming up with workarounds and fixes to product shortcomings so that the storage configuration bears at least some resemblance to the brochure;

• Sorting through the trade press for any information that goes beyond reprinted vendor press releases in the hopes of gaining a clue about what panacea technology the industry is preparing to spring on you next ;

• Adjusting management expectations about the performance they thought they would see out of all the shiny new hardware they bought against your advice;

• Trying to avoid technologies with built-in obsolescence so that the next fork-lift upgrade won't involve your cubicle; and

• Trying to figure out strategies for working in a bit of healthful REM sleep between all of the pager messages from your systems advising you that something has broken down or been broken into.

The continuing story of open systems storage is, for all intents and purposes, a quest to find ways to recover the management discipline and tools that were thrown out with the proverbial bathwater when centralized computing gave way to distributed computing models. The current drive to "utility storage infrastructure" is just the latest chapter in this story: an effort to go back to some earlier time when storage was centralized and virtualized and simplified ”and managed by folks who were professionally trained to do the job.

The problem is, however, that the storage industry, for all of its platitudes and industry associations, doesn't really seem to want a disciplined, well-managed, "storage utility" ”not if it means that a vendor will have to share floor space with a competitor's gear. To paraphrase the old reggae song, Everyone wants to go to heaven, but no one wants to die to get there.