White Paper: Selecting optical storage

Recent developments have made some optical devices very affordable, even for the smaller organisation, with genuine improvements...

Recent developments have made some optical devices very affordable, even for the smaller organisation, with genuine improvements in productivity and efficiency

With the increasing complexity and storage requirements of applications and networks, a viable alternative to conventional hard disk storage is often required to cater for these demands. As well as this, many clients are also looking for electronic storage and distribution of data rather than the more conventional format of paper. To add even more demands to the storage requirements is the fact that many applications now incorporate not only full colour images which may take 10 to 20 megabytes, but also large databases, scanned documents, and emerging multimedia and animations which are only one step away from a virtual reality environment.

However, storing and delivering this information now requires a more innovative use of available technology to satisfy customer expectations as well as improving productivity at the source of this information, which is not an enviable task for IT professionals. This is due to technology mispositioning as well as misconceptions that exist as a result of vendors failing to supply the negative aspects to a particular technology as well as the benefits.

Probably the most common technology that many offices currently use is CD-ROM, especially in the area of applications software and reference jukeboxes for images. One only has to install a complex software package from floppy disks to begin to appreciate the benefits of a storage medium where a single disc can hold up to 650Mbytes of data! But the question remains - what can these new technologies actually provide in addition simply to distributing application software?

The different approaches need to be addressed to appreciate fully what a particular subset of optical storage will actually deliver, not what the vendor promises. Indeed, one of the historic problems has been that a single technology vendor will try to broaden a product's appeal to a wider audience, simply because that vendor cannot supply the best solution for a particular problem. Without understanding some of the fundamentals, the user has little option but to go along with a lot of promises encompassed in technical jargon which does little to instil confidence.

The most important area to explore is how to increase the secondary storage issue so that client data which is inactive (completed projects) can be moved to a more economical medium, therefore freeing up valuable primary hard disk storage. The first essential is to try and adopt a medium that is a "computer peripheral" type of product. This generally excludes recordable CDs which will be looked at in more detail. This leaves either magneto-optical or phase-change optical drives as the obvious choices. These drives behave in much the same way as a conventional storage peripheral with data sectors being addressable individually for both reading and writing. The primary difference is that magneto-optical drives require two revolutions of the disc to rewrite data, whereas phase-change rewrites in a single pass.

However, as with any computer storage device, do not rely on any single medium for both backup, archive or even live copies of the data. No storage medium is infallible and adequate backup/copy procedures should be followed. An additional point to consider is that many optical drive vendors have now taken away the default option of recording data with data verify enabled. Naturally, this increases perceived performance, but what happens if three years later the data is unreadable due to, say, a media flaw that was undetected at the time of writing?

All this apart, placing an optical drive at either the client or server level will provide what is effectively an unlimited extension of the system storage, as the medium is removable. For large organisations that may require more online data, the natural extension is to provide an optical jukebox which places all the data effectively "near-line" with retrieval times of 10-20 seconds if a disc needs to be moved into a drive within the jukebox. Backup of this data can be either to a second drive, or even to a DAT (Digital Audio Tape) or DLT (Digital Linear Tape) system at the tertiary storage level. There are many software solutions from companies such as Seagate, The Cheyenne Division of Computer Associates and Q Star, to name a few, that can transparently manage this data migration process right through the primary, secondary and tertiary levels with user parameters to controlling the data staging process. Interestingly, the optical drives can, in many cases, use both re-writeable and write-once media with such software, and these discs can be mixed in a jukebox system so that more effective use of the technology can be exploited depending on the user's application demands.

Many system administrators may well balk at the cost of some of these solutions, but recent developments have made some optical devices very affordable, even for the smaller organisation, with genuine improvements in productivity and efficiency.

Many large organisations now implement optical technology to address specific storage management issues. One of Germany's largest publishing houses uses WORM discs to archive data in jukeboxes using Hyperdoc (from Dr Materna), and once the data is no longer required for access, the media is exported from the jukebox and can be used in a single drive if required at a later date. Interestingly, there are now many active and successful developers of software in Europe who write applications to optimise the use of optical storage within their data/image management systems. This naturally brings the benefit of "localisation" as regards languages. Even at the low end, PD (Phase Dual) drives which cost in the region of #300 are used extensively, and in the Middle East a large bank has adopted this technology to backup data on client PCs. The list of world class organisations using optical storage is extensive, and whilst a few may not have had the overall success levels expected, this has usually been due to incorrectly identifying the application and associated storage requirements.

The headlines and advertisements in the computer press have been full of CD Recordable as being everyone's dream solution, but does it really offer all that seems to be presented? The very idea of producing one's own CD-ROMs seems attractive with many possible uses, but in general the answer must be no. However, CD-R (Compact Disc Recordable) has some excellent characteristics that make it suitable as a companion product for existing storage solutions once the technical limitations are really understood. Data can only be recorded in "sessions" where the data must be kept flowing to the drive at a constant rate. Any interruption in this data flow will cause errors and render the disc unusable. Also, the medium is unprotected and, as such, is susceptible to handling issues. Due to its sequential "session" nature, there is no data verification "on the fly". The only way to verify data is to read the disc after it has been recorded and if any errors are present then another disc will have to be recorded.

But CD-R is a publishing medium, and a very good one. This is where clients can have data distributed in an economical manner on a medium that has wide industry standards at both the physical and file level. Essentially, CD-R is an electronic version of the more conventional, but cumbersome, laser printer or plotter. Apart from the obvious applications, new areas are emerging such as publishing intranet data where clients may not have Internet access. Ideally, if CD-R is the medium of choice, then an application needs to sit directly in front of the recorder to shield the customer from the unfriendly nature of the technology.

Software drivers which control the storage devices play as important a role as the actual hardware itself, so consideration needs to be given to the overall integrated solution, not just the individual components. Interestingly, in the case of CD-R, some software can more than quadruple the expected recording time of a CD due to its data preparation time! Of equal importance is the position within a network of the intended storage device. Where jukeboxes are concerned, it is usually beneficial to create a specific archive, or backup server. This then removes the potential dangers of introducing third party software drivers to an installed and functioning system.

With the correct match of appropriate technology to the specific storage or application problem, the rewards will be seen quickly both in improved productivity and service.

Compiled by Geoff Marshall

Read more on Software development tools