Is it possible that solids will ever surpass spinners in terms of areal density? I want to discuss this as evolution of current manufacturing methods. I'm not interested in hearing about some new magic material that is in experimental labs in some obscure university. Let's stick to realistic near-term evolution, say the next 10 years..

It would seem that we could pack enough NAND chips into a standard 3.5 drive housing; so many that 3 or 4 TB could be easily exceeded. And we might achieve 5 TB with spinners in the next year.

I see advancements on both fronts. Solids will always need some sort of extra internal wiring in the chip substrate itself to access the storage elements. And Spinners will always need a motor and some sort of read mechanism that needs to be moved around.

I'm not too concerned about speed here, both classes of devices are plenty fast as it stands right now.

_________________
On a clear disk you can seek forever.

Second thought -- scratch the speed comment.

Why are heads not independently positionable? Why must they always move as a complete assembly? Why must only 1 be active at a time? Is it manufacturing issues? Reliability? Complexity? Is there no need for such a configuration? Or is this sort of thing in serious development right now? Perhaps mfgs are sitting on it for future use? To me it would seem that if we have independently controllable heads we could seriously increase data throughput on Spinners. It would be like having a RAID-within-a-drive so to speak. Raid 0 specifically, as there'd be no reliability or redundancy improvements here.

And when I say that, I'm referring to gross head movements, not the piezo elements used as secondary actuators for fine positioning over a specific track.

Surely with multiple heads moving in different directions simultaneously we could double or quadruple the throughput quite easily. Each head attending to its own file or parts of a much larger file all at once.

I know they tried something like this in the 90's with CD roms but found it faster and more reliable to just spin the disc faster. Yeh spin it faster till it exceeds 60x and then explodes! That's the way to do it! And besides they did it with 7 parallel beams that weren't independently controlled. The scheme I'm talking about would be. With the reliability and capability of micro controllers today I don't see why this isn't done?

Finally, GMR heads seemingly read and write with a frequency of a little over 1GHz. Is there any technology in pipeline to increase this? What about the heat assisted magnetic recording tech coming out? Surely this will need to operate at a slower frequency and generate a lot of genuine British Thermal Units that need to be dissipated!

Discussion please!

_________________
On a clear disk you can seek forever.