erm... don;t all heads read and wright at the same time ?

they must do because the bits are spread across the platers, so bit 1 to platter 1 and bit 2 to platter 2..... if you could only do one at a time then youd only be able to write 1bit per revolution of the disk....

also.... bah, all disks are pretty reliable these days so Im going to say theres f all between them tbh

Quote from: Serious

Quote from: bytejunkie

wouldnt 4 platters be better? more heads to read so quicker delivery of data and can write more at the same time.

or is there a massive flaw in my understanding of disks.

Matt

IIRC only one head reads or writes at a time so no. The higher density gives a higher read speed at a given rpm.

More mechanical complexity equals higher likely failure rate. Higher density doesnt mean a higher failure rate at all. The tolerances in the disk are higher, but that doesnt change anything else.

I call BS.

Higher density = more likely to fail due to the nature of how bits are stored on a hard drive.

More bits per area = higher likelyhood of cross talk, or corrupt data.

Strange how density has gone up massively yet reliability has improved no end.

erm... don;t all heads read and wright at the same time ?

they must do because the bits are spread across the platers, so bit 1 to platter 1 and bit 2 to platter 2..... if you could only do one at a time then youd only be able to write 1bit per revolution of the disk....

also.... bah, all disks are pretty reliable these days so Im going to say theres f all between them tbh

The data goes right around the platter in a ring, the whole ring is read sequentially if needed. Unfortunately the data isnt spread across the platters (like the disks of a raid array) so there would be no point in parallel reading or writing.

Quote from: M3ta7h3ad

Quote from: Serious

Quote from: bytejunkie

wouldnt 4 platters be better? more heads to read so quicker delivery of data and can write more at the same time.

or is there a massive flaw in my understanding of disks.

Matt

IIRC only one head reads or writes at a time so no. The higher density gives a higher read speed at a given rpm.

More mechanical complexity equals higher likely failure rate. Higher density doesnt mean a higher failure rate at all. The tolerances in the disk are higher, but that doesnt change anything else.

I call BS.

Higher density = more likely to fail due to the nature of how bits are stored on a hard drive.

More bits per area = higher likelyhood of cross talk, or corrupt data.

Strange how density has gone up massively yet reliability has improved no end.

Quote from: knighty

erm... don;t all heads read and wright at the same time ?

they must do because the bits are spread across the platers, so bit 1 to platter 1 and bit 2 to platter 2..... if you could only do one at a time then youd only be able to write 1bit per revolution of the disk....

also.... bah, all disks are pretty reliable these days so Im going to say theres f all between them tbh

The data goes right around the platter in a ring, the whole ring is read sequentially if needed. Unfortunately the data isnt spread across the platters (like the disks of a raid array) so there would be no point in parallel reading or writing.

No serious... its the error correction thats improved with regards to data storage (alongside other tech like shakey mentions the hitachi perpendicular method of writing data) . Most data read from your hard drive nowerdays has corrupt aspects to it, rather than concentrate on writing it perfectly, they concentrate on bettering the error correction algorithms used.

Run spinrite on a modern drive, watch the ECC Corrected value from the smart readout. Youll be surprised.

Strange how density has gone up massively yet reliability has improved no end.

Id be inclined to disagree with this, albeit from my own experience and not working in the industry or anything like that.

"Back in the day" it seems hard drives barely ever used to fail, to my memory they just used to soldier on with no bad sectors and fire them up five, ten years later and they would work fine. I must have used hundreds of abused old office surplus hard drives growing up on various machines and tinkering projects and never ever had a hard drive fail once. More recently Ive had two go in as many years and you seem to hear about people RMAing hard drives quite a bit, when I was younger it was pretty unhead of.

Of course I am talking from my own limited experience but it seems logical to me that as they cram more and more data onto the drives, increase the spin speeds and just generally increase complexity of course drives are going to fail more often.