Holographic storage looks great to travellers in the storage desert thirsty for more archiving capacity. A 300GB drive to be announced later this year by InPhase could be the oasis we're looking for. There is a slight worry about its niggardly 20MB/sec transfer speed but this is a version 1 product so we needn't worry too much. Or should we?

Perhaps we should.

I talked to Bob Longman, the engineering director of Plasmon, about holographic technology. Plasmon is currently the largest blue laser storage media drive producer in the world and its 30GB capacity UDO optical drive is the market leader in professional optical archiving. There is a roadmap encompassing UDO 2 at 60GB, UDO 3 at 120 GB and a probable UDO 4 at 240 GB capacity. None of these are going to need holographic technology in Longman's view. After that, may be, but it is not a given.

Longman listed some facts of holographic storage life. These suggest that the InPhase technology won't become a reliable and commercially-affordable product for several years.

Facts of holographic life
For example the drive cabinet's length is currently almost twice as deep as a rack unit; it would stick out of the back a fair distance because it just won't fit in a standard rack space.

Then the actual media have data recorded on them and read from them by ordinary light. It's tremendously and accurately focused but it is ordinary light. If a media platter was exposed to the light before being fully written too then it would be like a piece of photographic film exposed to the light. Flash - and it's all exposed a uniform white and now useless. The holographic recording media is photo-sensitive.

That means the media has to be stored in a light-proof cartridge, unlike CD, DVD, UDO, Blu-ray and HD-DVD media. It also means that the drive has to be light-proof. Otherwise, when the cartridge is loaded and its cover opened to insert the drive's read/write head, light will flood in and expose the un-recorded areas.

Then the way InPhase puts more than one hologram in the same spot requires one of the light beams to be shone onto, or rather into, the recording media layer at an angle, a very precise angle. If you are going to have 100 holograms in one spot then you need the light beam shone at 100 different angles into the media, That requires an extraordinarily precise mirror mechanism.

What is more, it also requires the platter to be held in a fixed and very stable position. If it moves then the light beam's angle is wrong and the data can't be read or written properly. It is relatively easy to hold a spinning platter in a rigid position but the mirror assembly and the head have to move and such motion means they are more susceptible to vibration. Any vibration throws the light beam off-target.

Light power
To add more uncertainty to the holographic technology it is by no means clear that the writing of data to the drive will be simple. According to my information the sensitivity of the media varies over time. You need different light source power levels depending upon the age of the media. It may only have, for example, a three month shelf life before it becomes useless. That means customers cannot buy media far ahead and may also have to store bought media under stringent conditions, perhaps analogus to infra-red photographic film.

Furthermore the light power needed to record information on the holographic media varies after the first recording exposure and also with the sequence of exposures and the time since the last one. It seems that the drive will need to be aware of the exposure history and age of the media before being able to compute how much light power is needed for the next recording exposure.

This means that the history of a piece of holographic media needs to be associated with it, possibly be being recorded on the media.

It also raises the question of how easy it will be to move holographic media between individual drives and a question about the general interoperability of media and drives.

Net effect
The net effect of all this is that the current InPhase holographic drive assembly is technically very complex, quite heavy and rather costly. Combine that with its current size and we might conclude that what we are seeing is not a product, not a commercially available product, at all. It is more of a technology demonstrator.

There is another fly in the holographic ointment. No one has yet showed a way to mass-produce recorded holographic media in the same way that mass-replication of, for example, DVDs is possible.

It is likely that holographic storage technology will produce commercially affordable and reliable drives and media some day. But not in 2006, nor on 2007 and possibly, probably, not until 2010/2011.

It is my personal view that Plasmon will probably use holographic storage technology at some stage. Possibly it will be based on Optware's collinear technology rather than InPhase's multi-angled approach. The Optware technology conceivably has fewer technical hurdles to overcome

Whichever approach is used it will not be a commercial reality until five or more years have passed, and probably more. Holographic storage is the oasis that the thirsty traveller in the archive desert thinks he sees ahead. For the moment it is, like a hologram itself, nothing but a mirage.