
Seagate Execs
should be developing cutting edge SSD's instead of clinging to platters. The hybrid stuff is a decent middle step, but don't bank the future of the company on it long term.
Seagate's 4th quarter results beat expectations, but profits are substantially down and cost pressures rising. It beat the Street but the outlook is bleak. For its fourth fiscal 2011 quarter, Seagate's revenues amounted to $2.86bn – 7.5 per cent higher than the $2.656bn recorded a year ago. Net income went aggressively in the …
SSDs will not replace HDDs within the next TEN years - FACT!
The world wide capacity for creating storage on platters DWARFS solid state - FACT!
Solid state fabs are way too expensive to build, in the vain hope any driving down GB/$ - FACT!
Stop whinging about HDDs, buy an SDD (SLC if you can afford it) and dont keep data on it without a copy on your HDD and even better on a RAID based service somewhere.
I really don't expect to see Terabyte SSDs for £40 using flash memory technology. Not ever. Using memristors, a decade or two in the future, maybe. Flash chips are already quite close to their physical limits. In contrast there are a couple of platter technologies that are already working in the labs, that can give a 10x, maybe a 100x, increase in areal density. (HAMR and patterned media).
I'd have thought the problem for Seagate is the mousetrap problem. They can build a better mousetrap, but will the world beat a path to their door? Do enough people have use for a 10Tb drive? For a 100Tb drive? Progress in IT will ultimately be limited by economics before it is limited by physics, as the economies of scale start to fade.
If I were at Seagate I'd be putting a lot of work into hybrid drives - Gb of flash memory write-through cache, with Tb of HD behind it.
I'm already using this at home (media library). The biggest problem is access time - but a SSD cache on the side helps a lot.
I'm still waiting for HDDs and SSDs which have enough energy reserve onboard to commit or freeze their write cache if the power gets pulled. It's not as if supercapacitors to do it are that expensive.
in the case of journalling filesystems you really do benefit from the drive responding with a definite 'YES! th edata is on the media!', this is done with 'forced unit access' and 'flush' commands. essentially you do have a degree of control, but you need redundancy in the filesystem and a rewind point.