Vigilant said:
... I agree about slow hard drives. The cheapies buy 5400s, but 7200s are fast becoming the standard. And next year we should see some 10ks coming around. Most people don't realize that the HDD is the slowest part of their computer, the weakest link. ...
I agree.
Well, I don't know if these exist already, but I thought of more ways to make speedier drives (in addition to making them spin faster).
One would be to build them with two independant head assemblies (placed at 180 degrees of each other) instead of the one head set as in current drives. Any piece of information would at any time be 90 degrees from the nearest head on average instead of the current 180 degrees.
I see this as performing somewhat the same as raid 0. So, one would need only two of these drives to achieve what raid 0+1 provides with 4 drives. Firmware improvements would of course be required so that information is put on the drive(s) in such a way as to maximize the use of the "redundant" R/W heads in some way.
While we're at it, then why not also have heads that are capable of reading/writing more than a single track before it becomes necessary to move the head assembly, a bit like multi-channel tape recording equipment. I understand that this may not be easy to achieve without making the heads heavier and slower... But combined with two head assemblies, smart firmware could possibly more than make up for it.
My point is that there are other avenues to improving the speed of hard drives that seem unexplored at present, at least in currently selling products. The past several years we've seen hard drives improve in the areas or areal densities, rotational speeds, head miniaturization, head seek times, cache and host interface improvements. The time may be right to innovate with new features instead of improving by refining existing technologies.
I mentioned required improvements in firmware. Actually, depending on the intended use of the drive, external software (in the form of drivers or state altering utilities) could make these improved drives behave differently in order to optimize either the transfer rate or the access time (or balance between the two).
Has anyone heard of new drive designs such as what I'm proposing? Any comments?
As for optical storage, I've recently read an article about a new medium whose size is the same as a regular CD but instead of using reflection it would use fluorescence. One such disk could have tens or a few hundreds of recording surfaces accessible by changing the focus point of the lens, giving a storage capacity of about 1 terabyte. It seems like the research is completed on this and the first shipping drives, which would also be able to use regular CDs and DVDs, could be in the 100$ to 200$ range now if this technology were adopted. I don't remember about the other specs for this new storage medium. Anyone heard about this? I don't know about the expected price for the new medium itself.
Regards, ../Klingon
PS: I forgot to mention that as far as I know all heads in HDs currently move toghether. If they could move independently then each head (or each pair) could move faster and there could be more platters in drives without any reduction in seek times. Fragmentation would also be somewhat less of a problem.
I guess I should at least put a copyright on these ideas, if not get a patent. It might have to come to that one day.