The TAPE WARS Ain't Over, Folks!

Although Computer A Technology Review has been treating competition between tape drive formats and tape drive manufacturers as a year long war, the truth is that the war is an ongoing event, fought on multiple fronts, with results that reflect technology limitations, as much as they reflect how well perceived products or approaches meet the real or anticipated needs of the VARs, VADs, integrators, IT managers, end users, etc., who are selecting or purchasing them. If the previous 70+ word sentence was a bit long and convoluted, the next one should state the matter more succinctly. The Tape Wars ain't over yet, folks.

It's obvious that tape drives have been providing one primary benefit--the feeling of security--the perception that data is safe somewhere and can be restored, if there is ever a need to restore it. Fortunately for many, this perception is one that rarely must be validated--the number of catastrophic system failures that make such heroic efforts as full system restores necessary (and, thus, test the assumption) are infrequent enough for most users (present author excluded) that the ability to restore a system from one or more tapes isn't tested.

The actual reality is one that's not always realized. There's a story about a user who faithfully backed up data, on a regular schedule, until disaster struck. Only when he tried to restore the data did he encounter a problem--the media had worn out and was unable to store data. Although the backups were done, the data wasn't verified and a restore from tape wasn't possible.

Adherence to a regular rotating schedule of tape exchanges, offsite storage, drive cleanings, etc. can help avoid this type of problem. In many cases, the readers of CTR have been using such methods for years. In many cases, automated tape systems take care of many of these issues.

The backup software business has been consolidating over the past few years with the number of backup software vendors shrinking. Conversely, the number of tape storage devices has been growing with new drive models (Super DLT, Mammoth 2, Travan 5, AIT2, SLR100, DDS4. and others) and completely new drives (OnStream, Benchmark DLT-1, and Ecrix VXA) further complicating the picture. The new drives require software drivers in order to work with the backup software and backup software developers struggle to produce drivers for each new drive for all operating systems and hardware platforms that they intend to support.

Even if all tape software recognized and supported each new tape drive or automation product, selecting a backup software product is something akin to a marriage--once the software and hardware are matched, switching to a new software product can be a painful experience. Not only do the backup products create unique tape formats, these formats also make it difficult to switch to another product. A tape written on a drive, using one backup package will probably be unreadable using the same drive and another software product. Also, if it can be read at all, the new software package may not be able to write to the tape. Thus, being able to read data from a tape that has been written using a different software product may be only for the convenience of copying the data to a different drive or onto a hard drive for storage onto a new tape, written in the new software product's format.

Hardware compatibility can also be something of a problem. For example, Benchmark's DLT-l drive can read tapes written using a DLT4000 drive, but write tapes in a different format. Thus, a DLT-1 can be used to more rapidly restore a system that was backed up using DLT4000 drives or a pair of such drives can be used to copy and convert from DLT4000 tapes to DLT4000 cartridges written by a DLT-1 drive.

Often, newer generation drives feature read, but not write, compatibility, again making it possible to restore from an earlier generation's media or to convert from typically lower density media onto the current generation's media. Some newer generation drives also allow the user to take advantage of the higher speeds of the newer generation drives. For example, Sony's AIT-2 and DTF-2 drives can read the AIT and DTF media at the new drive speeds, delivering better performance than the drives originally used to write the tapes.

The point of all of the above is to make it clear that the Tape War is not simply about choosing the next, great new technology--there are a lot of other factors influencing the tape buying decision.

Tape formats last a LONG time. For example, there are more than a dozen companies currently shipping 9-track tape products, although 9-track has been out of the mainstream for decades. Legacy data, written using legacy drives, onto legacy media, may still need to be read--and there are still opportunities for businesses that support this need.

Selecting a tape format can become a very long-term decision. A decision today may well ripple through an organization for decades.

Choosing an end-of-life product or a product without a proven track record can be risky because future support may not be available and support infrastructures may not develop to support them. For example, in 1984, a tape product called Wafer Tape was developed for data storage. Perhaps a handful of people never heard of it. Today, there's no support for the format--it's probably nearly impossible to find the drives or the media. If a new format today is adopted by a small handful of users, the users may suffer a similar fate, if there's not enough market to provide future support.

A relative glut of new tape products and formats is streaming towards the hungry backup market. In many cases, formats will be competing directly with existing formats in price, performance, or capacity, or any combination of the above. Choosing a new format that delivers higher performance and higher capacity than a format that is already firmly established, but with inferior specs, may not be the no-brainer that it initially appears to be. There's much to be said for legacy devices--especially if they are already being used for regular backup.

Such factors as the size of a currently installed base (which suggests that a long term support structure should be available) and the number of drives installed (assuring fairly long term media availability) should be taken into account when assessing a move to a different drive technology. A technology roadmap that details multiple future generations of a particular product may be important to many users, who want a format that promises to grow along with the future storage needs, while also implying continual support (or, at the least, backward compatibility) with the current product.

A decade ago, for example, DDS was an important format with a developing, multi-generation road map. Buying and using DDS, followed by DDS-2 and DDS-3 drives, was a sound decision. Each new generation delivered higher performance and higher capacity, while also providing a useful level of compatibility. Today's DDS-4 appears to be the end of the line for most vendors. Sony has all but announced that it won't be developing a DDS-5. Seagate is also rumored to have stopped developing a DDS-5 and Hewlett-Packard appears to be the only company pushing the current DAT technology toward DDS-5. Although selecting DDS-3 or DDS-4 as a technology for backing up a new installation may not be a bad choice because the format is firmly established and media and support will be available, choosing a technology that has a longer term roadmap and a better likelihood of meeting its future targets may be a better choice for a new backup environment.

A Moot Question?

The question of which tape drive technology to select for backup may become somewhat moot for many organizations during the next few years and with the increasing band-width available over the Internet. The dropping price of hard drives may also have a significant effect on the desktop market. Tape drives aren't going to go away--but the issue of local tape backup may largely go away, if future trends continue.

In many organizations, desktop and notebook (often incorrectly called laptop) computers are backed up over the corporate network onto large corporate hard drives. Older data is backed off the hard drives onto tape. Once a complete image of the user desktop or notebook is taken, only changes need to be regularly recorded. These delta files, frequently containing only new documents and similarly small files, are stored as incremental change files and typically don't take up significant amounts of storage space.

When a user logs into the company network with his or her notebook, the network backup program can store the delta files in the background. In many cases, the files can be synchronized to the user's desktop computer. The point here is that, whatever device the user is currently operating, it's synchronized with the other devices that store the data. For a synchronized system, with complete programs and data residing on mutual devices, it may not matter much if your notebook is stolen or if a company fire destroys the desktop, as long as a complete, recently synchronized copy is on another device somewhere else.

Ubiquitous synchronization, with an additional storage/backup system also at work, will tend to make the issue of dedicated backup somewhat extraneous from the perspective of many users. Such software products as NovaStor's NovaNet-WEB or Storactive's Storactive already provide much of this functionality.

Third party organizations already offer storage and synchronization capabilities that can be accessed over the Internet. Similar services are available over corporate Intranets, using the products already mentioned and others expected to be announced in the near future. Remote backup for user systems, applications, and data will probably move the need for tape storage away from many organizations, onto secure tape farms. If this scenario is played out to the end, the Tape War may actually be fought by a handful of very large organizations, using extremely high capacity, high performance tape devices and massive amounts of disk storage for rapid response of most recently accessed or stored data. Although it is something of sacrilege coming from an editor at Computer Technology Review, it is possible that the days of the desktop or small server tape drive and tape arrays may be numbered, except as redundant storage that provides a comfort level that no external, unattached service, can deliver.

Of course, all bets are off when considering the use of tapes for things other than desktop and server backup. Secure, reproducible data storage for business critical data may always reside on company drives. Transaction processing, e-commerce, order processing, medical imaging, and other types of data may always reside on ever-larger tape drives. Digital video may always be stored and played from the (possibly less costly than hard disk) tape.

Tape won't go away, but it may move ever further from the desktop and from applications where system backup is a primary goal.

Backing The Desktop

The desktop, long a principal user of tape for backup, will soon have a different solution from Adaptec. For months, it has appeared that the cost of a hard drive and that of a tape cartridge with similar capacity are close to convergence. If you throw in the cost of a tape drive, in many cases it's less expensive to get a hard drive with the same capacity as a tape and tape drive. A hard drive provides random access and better performance, but considerably less portability, and, for most tape formats, much more sensitivity to shock and vibration.

A hard drive isn't the same as a tape cartridge. There's room for both.

For the desktop, Adaptec has smudged the distinction, delivering a product that uses a hard drive to deliver disaster preparedness and recover. The product, called reZOOM was scheduled to begin shipping in mid-November.

reZOOM copies selected data or entire drives from the desktop drive(s) onto a dedicated hard drive that is installed into the user's desktop. At regular intervals (that can be predetermined by the user), delta files can be copied from the user's drive onto the reZOOM drive. The reZOOM drive, thus, stores a frequently updated, complete image of the user's drive(s).

If the user's primary drive should die or files become corrupt, the two drives can be compared and reZOOM can automatically restore the drive to its pre-damage state. Unlike such products as GoBack, which operate in a somewhat similar manner, taking regular system snapshots and allowing for recovery to a selected time and date, reZOOM uses a dedicated drive that is not designed to be modified by the user and supports both SCSI and ATAPI drives.

The product is pretty amazing. In addition to actually being able to restore a damaged drive, reZOOM can be used to write the contents of the reZOOM drive onto a new hard drive. For users looking to upgrade from one hard drive to another, reZOOM can make a complete copy of the original drive onto a new drive. If, for example, you wanted to upgrade from a 9GB hard drive to an 18GB drive, reZOOM can restore the system onto the new 18GB drive and can also automatically create a full 18GB partition.

Additionally, because reZOOM stores snapshots of the drive and files, it can be used to recover files that may have been erroneously deleted from the hard drive. What reZOOM does seems to be a lot like backup. With the price of hard drives continuing to plummet, a dedicated drive with reZOOM installed may be an attractive, low maintenance form of system backup.

reZOOM can automatically generate e-mail when a drive fails. In theory, this can be useful for contacting a vendor's technical support department, allowing for the vendor to proactively assist the user in recovering the damaged system. The product will initially be available as a shrink-wrap product, although Adaptec is exploring OEM agreements that will bundle the product with hard drives.

A major limitation is that, if the computer is stolen or damaged, reZOOM is of no value in recovering data. For that, there's tape and the remote synchronization products already described.

In this, perhaps strangest of the Tape Wars articles, I've explored some of the reasons why the war may be changing. The question of which format or whose drive will emerge victorious may be considerably different a few years from now than it is today, if the factors described here come into play.

Remote backup, synchronization on multiple devices, and products like reZOOM that store drive images onto a hard drive installed in the desktop may lessen the need for tape for many users. While the need for backup and restoring will always exist, the question of where the backup will be performed and who will do the backup may continue to change.

Business critical data will probably always require tape backup. Electronic entertainment services may always have to store much of their programming onto tape. The amounts of data that are stored onto tape will probably continue to climb at enormous rates. The devices that store these gonzo amounts may also change, evolving in a way that makes the current combatants in the Tape Wars look puny.

However, the need for backup continues today. The practical reality is that, for 1999, 2000, and at least a few more years, the current crop of competitors will continue to be increasingly significant. There is no clear winner. There may be multiple winners, serving multiple markets, and the landscape will continue to change.