The decision of Cipher Data Products Inc’s Optimem unit to cancel introduction of its erasable optical disk drive (CI No 735) highlights the fact that the biggest problem facing the optical storage industry is the lack of standards. What with widely differing error recovery schemes on disk drives, varying sector sizes – take the manufacturer who implemented 7,904 bytes per sector as an extreme example – not to mention the de facto SCSI interface standard being treated differently depending on manufacturer, a lot of people are blaming each other for the industry’s present state of play. Ken Cross, vice president of systems at Perceptics Corp, Knoxville, Tennessee, has some hard things to say on the way hardware manufacturers do business. Cross, whose company deals in optical storage development, pin-points non-standardisation as one of the biggest drawbacks as far as software development is concerned.
Jukeboxes
Every disk drive manufacturer has a different code for errors and we then have to support each individual drive, he complained to Microbytes Daily. Consistency would be nice. He thinks it’s a shame non-standard measures are used by vendors who like to call them features, especially as they do so with all good intentions. He points out that non-standardisation of hardware not only affects software developers but hampers the industry as a whole. Computer vendors are slow about supporting the optical market because they are waiting for the technology to settle down, he explains. Well, it has settled down and there is no reason to wait any more. Cross calls for the help of optical media manufacturers in tackling the problems experienced by users of auto-load optical storage systems. He suggests they start adding encoded serial numbers to the media to cut down confusion, especially where jukeboxes are concerned. Volume name and shelf location are used to identify certain disks, says Cross, but for large systems with thousands of disks, this wouldn’t be practicable. What is needed is a unique identification number on the disk that is put there by the manufacturer. This way, an operator couldn’t enter the wrong disk ID. Memory consultant Richard Zeck, backs up Cross’ statements but claims that the lack of an industry leader is another cause of the optical storage industry’s problems. The very diversity of optical storage options is perhaps its biggest negative, he says. IBM’s announcement of its 3363 [optical drive in April] was very important, but we’re still waiting for the overall breakout type of leadership that’s needed. Speaking at a conference on optical drives in San Francisco last month, Zeck said we will see three, perhaps five, generations of products, and we are only in the first generation. He said by the 1990’s, optical storage will be a very large successful market. Within this market he sees several trends developing, particularly with the emergence of 5.25 half-height drives.
Holographic
There’s no question about it. A majority of the optical drive manufacturers will be making 5.25, and before long 3.5 units that are half-height because this is what the market wants. Zeck reckons manufacturers have been able to develop smaller drives with multiple heads because of the evolution of holographic optical elements, higher power laser diodes (which use shorter wavelengths and diode arrays), better position sensors, and single-element moulded aspherical lenses. Diode lasers with shorter wavelengths will mean bit and track densities can be doubled, effectively quadrupling storage capacity, and laser diode arrays will permit multichannel read/writes. These laser diode arrays may incorporate fibre optic arrays. Within the next three years Zeck predicts such innovations will be widely used. Access times are on the increase too with the market clearly indicating to manufacturers that 50mS isn’t acceptable. In the near future Zeck says we will see drives spinning faster and access times of 30mS. Despite all this he doesn’t discount magnetic storage altogether. He sees the continuing evolut
ion of magnetic technology, which annually improves in both capacity and throughput and decreases in price as one of the greatest challenges optical storage must overcome. Magnetic storage can still outstrip optical storage in certain environments, and systems such as Konica’s high-capacity 5.25 disk drive will go on make life tough for optical devices.
Multifunction
Content from our partners
As for that Optimem drive, Peter Tobias of BIS-Mackintosh Optical Storage Program in Luton, Bedfordshire, reckons that the company had no alternative but to put the drive on hold. It’s a multifunction 5.25 drive that is designed to use both an erasable medium and to read CD-ROMs. It cost a fortune to develop, but in the current state of the company, with only $20m annual turnover, it needs to concentrate on its write-once drive to establish a position in the market. The Optimem drive, he says, uses a magneto-optic technology, with medium developed in collaboration with 3M. And it is by no means clear whether this technology or the pure optical phase change approach, where the spots on the disk are switched by the laser between crystalline and amorphous states, will come out on top. Tobias points out that IBM’s published papers suggest that it is majoring on phase change, as is Philips. Despite the fact that people who approach the subject from the point of view of the data processing industry see erasable optical drives sweeping write-once drives into the dustbin of history as soon as the former are freely available to an agreed standard, Tobias forecasts a rather longer life for write-once technology. Erasable media are 50% more expensive, and the capacity of write-once platters is higher. Write-once technologies now have guaranteed lives of between 15 and 30 years, so that they can be certified for applications where there is a legal obligation for records to be retained for several years. As a result, it is not the comparatively short-life archival tape that they threaten, but the paper, microfilm and microfiche records that are still widely used for long-term records.
