Who needs NAND when rust never sleeps? Seagate dines out on nearline disk drive boom Seagate's top line number bounced and profits took off in its third quarter as it made the most of increased nearline disk drive demand. The big question for us, the big question for everyone [about the nearline market]... is how long is it going to last and how big is the peak? Revenues for Q3 ended 30 March were $2.8bn, 4 …
our crappy san with 18 nearline R10 2tb and 2 SSDs running as cache performs almost the same as our not as crappy 22 R10 15k 900's for our VMs.
both SANs are MD3220i running mpio to 2012r2 servers. only 9VMs but one is exchange and one an SQL server.
the NL san was half the price, even with the 2 SSDs.
The tech necessary to drag out tiny variations in the magnetism of a disc has been around some time. It is commonly used to recover data that has been overwritten and clearly demonstrates how a single bit of data on a hard disc can have more than just the two values of one and zero.
Now consider a hard drive with a second read head. In write mode, it reads ahead of the read/write head supplying the precise orientation of each bit in the write heads path, so that the power can be specifically tailored to turn that minuscule section of the disc from its current value to one that can be identified on a scale of zero to seven, or nine, or maybe even 15, so we could write hexadecimal code directly to the drive, where once we only written a binary one or zero.
A 2TB drive would then become a 32TB drive. Write times decreased by a factor of 16 and read times decreased by 32 as the two heads share the work.
Way back in computing's early history a couple of attempts to go bigger than binary ended in failure, due to the complexity and vagaries of early electronics. Those are no longer stumbling blocks, and I'm puzzled why that direction isn't being explored today.
"We would imagine that, with demand high, it can basically build me-too-and-a-little-bit-better product and grow its NAND business significantly."
On the other hand, buyers who've been held down over a barrel by Seagate and WD for the last 8 years WRT disk pricing and reliability (since the Thai floods) may take the opportunity to say "adios" to both vendors(*), given that it makes much more sense to go with a vertically integrated supplier of enterprise silicon storage than one which actually _let_ reliability fall and slashed warranty periods to the bone when it saw it could get away with it
The fact that most of the "vertically integrated suppliers" have long-term plans and are more likely to be around in a decade, (vs companies which are focussed on maximising shareholder yield per quarter even if it kills long-term customer trust) also plays into that equation.
There's no new HDD research coming down the line. The R&D labs are closed and {H/M}AMR is already years behind schedule.
(*) Yes, I know there are technically more than two, but Toshiba's share is basically just a few scraps sold off to keep the chinese competiton regulators happy.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
