Time and again, a predictable process has played itself out across the (bring up the reverb) Grand Canvas of History. Organizations form and become successful. Success breeds pride and arrogance, and, as the world changes, market forces and hubris bring about the organization's decline. Rome, Communism, dotcoms — all have followed that pattern. Does anyone learn from history?
Hubris brings us to this month's subject: mitigating chaos in the business environment. Entropy always wins, and that 'puter on which you crank out your production is destined for failure. It just hasn't happened…yet.
It's easy enough to make your computer wobbly and even fall right over. Blue Screen or Bomb, take your pick. Simply install lots of software and hardware without updating drivers or OS and never do a disk check when the computer takes a dump. A skilled IT bubba can stabilize NT or classic Mac OS and keep it running without incident for extended periods. Just don't change anything.
Installing new stuff is one major cause of unsavory behaviors on the part of your computer, and a simple check of the vendor's Website prior to installation will alleviate much wasted time and money. When things do take a turn for the worse, as they inevitably will, you need a disaster recovery plan. The key to even the most basic plan is a proven data-restore mechanism. I'll say it again until you can repeat it in your sleep: Who gives a rat's booty about a backup if you can't restore it?
Back in August, I talked about some value-priced tape offerings, including DLT1, progeny of Boulder's Benchmark Storage Innovations (www.4benchmark.com). The parent of DLT1, the DLTape format, premiered way back in 1989 from the late, lamented Digital Equipment Corp. during its MicroVAX II days. It had a whopping 2.6GB capacity. As time went by, the DLT family grew and Quantum (www.quantum.com) wanted to provide a small-form factor, low-cost option for the budget market without diluting existing sales. Enter Benchmark, which licensed Digital Linear Tape technology from Quantum and created its own product niche for the low- to mid-range segment of the tape drive market. In turn, Tandberg Data also licensed DLT technology, so Tandberg, Quantum and Benchmark now all offer DLT1 products in both manual feed and robotic changer configurations. Tandberg also has a competitor to VXA and DLT1, the very affordable SLR100 format.
The first thing, folks, to look for in backup hardware is speed and capacity. Native capacity for DLT1 is 40 GB, hefty enough for most small businesses. A 3MB/s native transfer rate won't frighten small children, but gets the job done nonetheless. Remember that, though newer linear formats like VXA double that native transfer rate, real-world throughput has as much to do with the host's abilities as the attached tape mechanism itself. If the host can't keep up with the tape, then nothing is gained by having a rip-roaring mechanism.
“Benchmark's DLT1 uses standard DLTtape IV media, and the DLT1 boasts backward-read capability with the popular DLT 4000 format.” So says their spin doctors, but notice the bit about the 4000. DLT1 can read DLTape 4000 tapes but cannot write that format. DLT1 is also compatible with virtually all popular backup software packages. So, if you're in the market for a fixed-head tape backup scheme and need the added bonus of 4000-read capability, then consider DLT1.
For those of you who have a commodious wallet and the need for higher throughput and capacity, there are a host of new options. Two recent entries, LTO and SDLT, are competing head to head in the high end of tape-storage solutions, alongside Sony's AIT-2 format and Exa-byte's Mammoth-2. Super DLT is the Clark Kent of the DLT line and Jor-El's, er, Quantum's bid to maintain DLT's predominance. SDLT provides a current native capacity of 110 GB and native transfer rate of 16 MB/second. It also offers “backward-read compatibility with multiple generations of DLTape 4000, 7000 and 8000 Series drives and millions of DLTape IV data cartridges.” In addition, non-BRC or “non-backward-read-compatible” drives are in the works for use in tape library automation applications.
The Linear Tape Open format was developed by IBM, HP and Seagate as a purposely bifurcated family of application-specific products. One branch is the Ultrium format, which strives for high capacity. The other is the Accelis format, which forgoes capacity for extremely low access time. Accelis drives achieve “sub-10-second average” search times, allowing tape to be used in online storage. This makes an argument for optical disc libraries more difficult to justify. Another LTO feature, borrowed from AIT, is the inclusion of in-cartridge, solid-state memory for access to TOC information without touching the tape.
The trade-off for Accelis is a current native capacity of 25 GB, measly in comparison to Ultrium's largest current 100GB native cartridge. Both formats share a 20MB/second native transfer rate. Never slacking, the designers are looking to the future with projected capacities rising to 200 GB for Accelis and 800 GB for Ultrium in the fourth generation.
Next month, I'll look back at the past century, about to slide off quietly into the history books. In the meantime, don't be lulled into a sense of false security. Always brush your teeth, get plenty of rest and, no matter what format you use, back up often!
OMas, in an effort to make sense of things, has enjoyed the outstanding sailing weather surrounding the tiny Pueblo By the Bay. This column was created while under the influence of Arthaus/Kinowelt Media's modern interpretation of Le Nozze di Figaro on DVD-V. Links and occasional commentary atwww.seneschal.net.
PEDANT IN A BOX
Many tape formats, whether audio or data, rely on physically fixed head assemblies, while others, such as video decks, use heads that move relative to tape in a rotary fashion. To achieve the high data throughput necessary for wide bandwidth signals, designers have to either move the tape really fast past the head or move the heads really fast past the tape. The former approach spawned SDAT, or Stationary Digital Audio Tape, formats like DASH and data formats such as DLT, SLR100 and VXA. The latter approach, pioneered by Ampex in their revolutionary video tape machine, employs a rotating head assembly that spins past slow moving tape. Video tape technology has morphed in data formats including DDS and Exabyte's family of products.
While data formats are moving away from the complexity and unreliability of rotary head designs, the majority of digital audio tape formats, including RDAT, ADAT and DTRS, are rotary designs derived from consumer videotape transports. The first RDAT was Sony's PCM-1600, a substantial beast that converted analog audio into linear PCM and then channel-coded the digital data as monochrome NTSC-composite analog video for recording onto a companion ¾-inch U-matic tape machine, another Sony format. They later created a consumer version of the professional transcoder, releasing the PCM-F1. The entire encoder and tape transport product was later integrated and packaged as the RDAT format, the first turnkey digital audio tape meant for consumers.
3M and MCI/EMI made early SDATs for the pro market but they never caught on. Later, Philips made an unsuccessful attempt at launching a consumer SDAT format, the DCC or Digital Compact Cassette.
Native capacity refers to the ability of most tape mechanisms to losslessly compress incoming data in real time prior to writing to tape. Typical lossless compression achieves only a 2:1 ratio at best. This compression, however, yields a doubling of tape capacity. Because the amount of incoming data is reduced prior to writing, twice as much can be shoved through the pipe as well, yielding double the throughput.
Tape mechanisms allow software drivers to turn hardware compression on and off via external commands. The amount of redundancy in the data determines the amount of compression. With compression enabled, backup software never “knows” what the total capacity of a particular tape is until it's full! So always go by the native numbers hidden in the fine print.