The Register® — Biting the hand that feeds IT

Feeds

Hitachi GST pushes out boosted SSD

Intel's new NAND

Cloud storage: Lower cost and increase uptime

Hitachi GST has birthed a boosted Ultrastar SSD using Intel's latest 25nm NAND. Intel launched its own 520 just a few days ago.

In these frenetic days of flash hyper-awareness, close attention will be paid to the performance of new flash drives to see if suppliers are keeping up with the pace or slipping behind. The Ultrastar SSD400S.B has jumped from Intel 34nm NAND to 25nm and gained a near 30 per cent boost in random read IOPS, delivering 57,500 of them versus the first generation SSD400S products 41,000. The other gains aren't quite so great but still good. Random write IOPS improve from 21,000 to 25,500; sequential read bandwidth from 516MB/sec to 536MB/sec; and sequential writing from 458MB/sec to 502MB/sec.

Intel's own 6Gbit/s SATA 520 delivers 50,000 random read IOPS, 42,000 random writes, 550MB/sec sequential write bandwidth and 520MB/sec bandwidth for sequential writes. Capacity points for the 520 are 60, 120, 180, 240 and 480GB – whereas the Hitachi drive, with a 6Gbit/s SAS interface, offers 100, 200 and 400GB.

Obviously performance improves generally as capacity goes up, but Intel's own drive looks a better performer than HGST's Intel-based drive. We know the Intel drive uses Sandforce SF-2281 flash processors with exclusive Intel firmware. Perhaps the HGST product does not use the Sandforce processors? It certainly can't use Intel's firmware.

The HGST drive is faster than the 520 on random read IOPS but slower on the other three main performance numbers, and this looks like Intel competing with its own channel and not doing HGST any favours. In fact it looks like HGST got the short end of Intel's flash stick. We can't imagine coming HGST owner Western Digital looking favourably on the HGST-Intel deal. After all, it could buy in Sandforce controllers too.

HGST gives its gen 2 400S drive a lot of endurance. The 400GB product can withstand writing 19.2TB a day for five years before starting to wear out; that's 35PB over the life of the drive. Tom's Hardware thinks the 520 can write 7GB of uncompressible data per day for 5.12 years, and it looks as if HGST has the endurance edge, although we don't know if its endurance is for compressible or uncompressible data, assuming the latter.

We're told potential OEMs are qualifying the HGST drive now and production should ramp up by the mid-point of the year. ®

Steps to Take Before Choosing a Business Continuity Partner

Latest Comments

What do we actually need SATA/SAS for?

It's all good and well if we still have a controller on the PCI bus between the RAM memory and the hard drive, as long as the hard drive is a spindle.

With SSD I don't think there is a need for a SATA or SAS controller, which architecturally slows down IO transfer to and from the SSD memory chips anyway.

Until we have the Exabyte storage chip solded onto the motherboard, SSD memory should be on PCIe cards.

Prove me wrong, in my opinion the days of SATA are counted, I guess.

0
0

I saw a few articles that said it was SLC, and just now found the official line: http://www.hgst.com/press-room/2012/hitachi-gst-ships-the-industrys-first-25-nanometer-slc-nand-flash-enterprise-class-ssds

I think you're right about the 520 only being MLC though - In that case, the price difference will be even greater. The 520 may well be suitable for enterprise use, in certain scenarios, (just as X25-M etc were, with over provisioning), but it's not really a competitor for this drive...

By successor to the X25-E, I meant from Intel...

0
0

Nah...

If this drive was SLC it'd be quite a lot faster. It's probably MLC in the Intel 520 (thus slightly higher performance) and eMLC in the Hitachi, for more endurance.

Also a major consideration in drive lifetime is the quantity of spare NAND available to use as blocks start failing - the 520 comes in 240GB and 480GB advertised capacities, whereas the Hitachi is available in 200GB and 400GB, with probably the same amount of raw flash on board, which would allow them to withstand ~20% more failed blocks before the user noticed anything.

"""so maybe there will be a true successor to the X25-E?"""

There are plenty of those - they're made by companies like STEC, are blisteringly fast, and far, far out of my price range. Just like the original X25-E.

0
0

More from The Register

SCO vs. IBM battle resumes over ownership of Unix
Zombie lawsuit back and wants to suck the brains out of Linux
 breaking news
What's HP got under wraps? Looks awfully flash and tape shaped
What happens in Vegas won't stay there - we've got the details
Microsoft borks botnet takedown in Citadel snafu
Stupid Redmond kicked over our honeypots, wail white hats
IBM's $1bn layoffs latest: Now axe swings in US, Canada - reports
Union claims 121 storage bods canned after dismal sales
NetApp musters muscular cluster bluster for ONTAP busters
Storage array OS overhauled to juggle more nodes, go down on you, er, less
HP adds 'Haswell' Xeon E3s to entry ProLiant servers
Gussies up MicroServer for SMBs, adds baby switches
Buffalo herds DDR3 RAMs into DriveStation's spinning rust corrals
Claims cache-packed gear keeps up with flash drives
'THINNEST EVER' spinning terabyte beauty slips out of WD fabs
Size-zero drive packs a whopping 143GB per millimetre