Linux recovery fat

One of the myths in the IT world is that most methods of backing up data are ineffective. Thomas Pynchon's 1961 novel "V" popularized the phantasmagorical notion of giant blind albino man-eating alligators roaming the sewer tunnels under the streets of the Big Apple. Likewise, the IT press has popularized a no-less sensational notion: that the vast majority of data recovery attempts fail. A recovery failure ratio of 70% is a recurring statistic apocryphally attributed to "analyst research." This means that no fewer than two out of three attempts to recover data from backup end in failure. Other statistics demonstrate that business managers and those whom they charge with protecting their business data uniformly lack confidence in the usefulness of their data protection practices.

The Pot and the Kettle

Faced with this--"ahem"--unpleasant perception dominating the customer base, storage hardware manufacturers and storage software vendors have allowed their marketing dialectic to turn into a finger-pointing exercise, which would be amusing if it were not so tragically ridiculous. Over the years, we've become familiar with the point-counterpoint of "It's the software!" coming from the hardware bench, with "It's the tape drive!" answering from the disk vendor's bench, punctuated now and then by "It's the user's fault!" coming from all vendors in unison. Fratricidal mud is slung daily between marketing mavens of rival companies. But does it really serve any useful purpose to portray a competing product as unreliable? Given the overcrowded state of the storage market, most competing products are indistinguishable to the customer's naked eye, and the perception of unreliability spreads beyond its intended target to stain the entire market space, including the accuser. The down-home country term for the kind of marketing messaging dominating the storage industry today is "spitting in the wind."

D2D: Proceed at Your Own Risk

Emerging from the rubble is a new paradigm which shifts the backup and recovery limelight to disk storage. In the late '90s, EMC Corporation's Mike Ruettgers was known to say to anyone willing to report the story, "Tape is dead!" Yet tape is still alive and well, in spite of his pronouncements. Even EMC has changed its tune to, "There could be a place for tape." Just recently, a senior executive in a market-leading tape storage company finally responds to Ruettgers' claims with, "I would argue that disk is dead." But the fact is that disk itself is not the biggest threat to tape's viability.

The real threat comes from bandwidth. The fat lady will sing for tape and other removable storage media when, and only when, it becomes economical and feasible for everyone and anyone to spirit data off site via wide pipes. Until then, relying on inexpensive consumer-class disks located on site as a substitute for proper disaster recovery planning is more aptly described as planning for a disaster. A predominant portion of the landmass of the U.S. is subject to significant natural hazards: damaging earthquakes are known to strike even in the Midwest, fires periodically scorch the West, floods and tornadoes repeatedly threaten the Heartland, and seductively-named hurricanes wreak seasonal havoc on the east and southeast coastlines. For these reasons, real estate lenders commonly require proof of specialized insurance from their borrowers. But, to date, business lenders do not yet require proof of insurance or protection for their customers' data assets, which are even more irreplaceable and valuable than buildings and parking lots. But just leave it to Uncle Sam to step in where the private sector's voluntary measures fall short. Big Brother's new legal morass of Acts and Regulations governing data retention and handling virtually guarantees that there is no reasonable or achievable path to compliance for the average small to mid-sized business, particularly not if D2D is its sole method of backup.

Disk Staging Takes Center Stage in SMB

Despite the risks, D2D backup has swept through the SMB space like wildfire. Like the television watchers in the seventies film "Network," SMB tape users have flung open their windows and stuck their heads out to yell: "I'm mad as hell and I'm not going to take it anymore!" Like Linux sweeping through the Unix space, D2D has gained ground as a grass roots movement because ultimately it makes sense to users. With the cost of disk storage dropping and dropping, it is difficult for anyone to argue against, at the very least, using disk as a staging step preliminary to migration to tape storage or other removable storage. Count the benefits: ease and rapidity of recovery, inexpensive high throughputs, multiple data streams, improved utilization of expensive removable storage resources, and the list goes on. And, by eventually migrating to tape, the user gets the benefit of secure off-site archiving, plus low-cost durable media for long-term retention of data. It's no wonder that Disk-to-Disk-to-Tape (D2D2T) is one of the storage industry's hottest buzzwords. Vendors of every ilk are doing all they can to encourage this propitious feeding frenzy. But who should take the credit for bringing D2D2T to the masses? Several vendors are vying for the after-the-fact distinction of being first in this new category. One vendor even went so far as to file a trademark registration application for the term D2D2T as recently as July 2003--months, if not years, after the term had already been in common usage. The truth is that the credit for D2D2T should go to the prescient users who envisioned a new and sensible way of doing things long before the legion of greedy vendors saw the potential of a lucrative bandwagon and jumped on.

[FIGURE 1 OMITTED]

Figure 1 shows the integration of D2D2T in the network

Virtual Tape: Another 70's Throwback

Even with users pioneering a common-sense approach, some vendors are still pushing anachronistic solutions. The old timers of the data protection world are today sitting in the back of the classroom snickering "Heh, heh, let's make the disk act like a tape." Whatever led IBM to introduce the notion of a Virtual Tape System may have been a reasonable tack to take in the mainframe-centric world of the '70s and early '80s. "Centralize, centralize," was the mantra of that age. Why not then centralize and coordinate backup streams on disk before transferring them to the high priced tape media available in that era. But, by any stretch of the imagination, tape emulation in today's world doesn't work, and is the result of collusion between market-leading entities whose primary purpose is to keep their customers and protect their own vested interests. This leads to tape vendors bringing to market a new generation of solutions that support wasteful, repetitive and ineffective procedures.

Butlers, Chauffeurs and IT Managers: Vestiges of Antiquity

Compared to 50 years ago, very few households have a full staff today dedicated to domestic tasks. Modernization and wide availability of consumer goods have done their share to obviate the need for servants to perform household chores. But even more, the rise in the minimum and expected wage has made it impossible for middle-income households to afford the luxury of servants. Likewise, in the very near future (if not today already), the average small business will no longer have an IT department. Today's improvements in Information Technology may or may not be leading to lower requirements for professional management. But the argument is moot, since the rise in wages and the continued trend in the economy make in-house IT managers unaffordable to most small businesses. Therefore, they must do without them. An entire generation of college-trained IT resources today faces flipping burgers or other menial tasks as its only career choice. Everyone has a story about an Oracle DBA doing janitorial work until things pick up. As much as one would like this to be a manifestation of a short-term cyclical trend, it is not. The job function of the IT Manager will be just as surely and permanently extinct in the SMB world as that of a butler or chambermaid in middle-class America. The resulting conundrum: How will SMBs manage increasingly complex technology requirements while their ability to afford the resources needed declines and even disappears? The answer, put simply: They can't, and they won't.

To Be or SMBee ... That is the Question