SSD Prices On Parity With High-End HDD By 2011

kgagne writes "EMC executives were heavily pitching the virtues of solid state disk drives at their annual users conference in Las Vegas, saying that SSD will not only be on price parity with high-end Fibre Channel disk drives by the end of 2010 or early 2011, but that NAND memory will solve all sorts of read/write issues created by spinning disk technology. EMC's CEO and its storage platforms chief said the company will do everything it can to drive SSD prices down, and adoption up, by deploying them in their products. One issue might be that EMC is using SSD from STEC, which is being sued by Seagate for patent infringement." The article also mentions some of the work EMC has been doing to make sure SSD is enterprise-class reliable, such as developing "wear leveling" software.

Comments

Re:The Future is Solid State

[hostyle]

-1 Re-Iterating The Damned Obvious

Re:The Future is Solid State

[kesuki]

"One cannot know whether or not the SSD makers "lie the same" as the disk makers."

Sure you can, they're Human of course they lie. MTBF can be generated, based on how long it takes for 'more than 50% of the data sectors fail' but who would keep using a disc that kept having randomly failing disc sectors, even if SMART technology can reduce the risk of loosing data...

so of course SSD makers are going to calculate a MTBF assuming the type of wear a typical person who boots vista once, for every time it crashes*, and does nothing but plays solitaire, freecell and spider solitaire all day! it takes a lot of hours of playing solitaire, a non disk heavy activity for a SSD device to fail!

remember NAND flash memory is based on changing the structure of a semi-conductor with an electric charge, the more you do it, the sooner the device fails... even reading the state of the material causes wear, because electricity is used to read as well as 'write' to the memory, kinda the way a laser erodes the dye on a recordable optical disc... obviously though if you can Write millions of times, you can read billions of times. there is another problem with NAND though, NAND memory is often made with tantalite, a rare mineral used in semiconductors and capacitors..

if we recycled 100% of computer parts, the tantalite problem would be solved easily, but we're not even close to 10% recycling... so even as we speak gorilla habitat in Africa is being destroyed for tantalite mines. they've been building more and more of the mines, since a price spike in 2000, and the number of mines running are keeping the cost of tantalite from spiking again, but it would be so much better if we just recycled our old computer waste. we could save gorillas, if we pushed for refundable deposits on recycling electronic devices like computers, etc. if it worked for the lead acid battery it can work for tantalite, copper, aluminum, gold, and silver in electronic devices. FWIW plastic in electronics can be recycled into diesel fuel, as i found out from here. http://www.cbc.ca/news/background/small-business/recycling.html

*= which is probably every day, more if you install hardware drivers.

Re:What about filesystems...

[kesuki]

The only adaptation I can see is trying to minimize wearing on certain blocks, but from the looks of it the SDD's are being designed with wear leveling in mind so I doubt even that will matter to the software. Actually with proper software you'd probably like to do the opposite - try to wear out certain blocks as fast as possible. This way the lossage is more predictable and rest of the disk is kept in a better shape. Point being that bad sectors aren't really a big deal if you're prepared for those. When NAND memory fails, it can fail in such a way as to make the ENTIRE flash memory device unreadable... this is from real world NAND memory devices failing from real world use, all of a sudden not wear leveling seems like a suicidal mode of wear... if the entire chip can short out from a single block failing.

"In case of a massive damage,

        * If the device is not accessible at all (circuitry failure), no software can even attempt the recovery. Physical intervention is required.
        * Even if the device seems accessible, software recovery run the will take excessive time to complete, making the attempt impractical. On top of that, the recovery run puts further stress on the device. This may be undesirable.

In case of the massive damage, there is no point in attempting the do-it-yourself type data recovery at home. There is little you can do to repair a physically damaged device without the special equipment. If you have a physically failed storage device, we have a discount available for a DriveSavers recovery service. DriveSavers are quite good with physically damaged devices and we recommend you contact them if need arises."

http://www.z-a-recovery.com/physical-flash-memory-failure.htm

Re:What about filesystems...

[kesuki]

When NAND memory fails, it can fail in such a way as to make the ENTIRE flash memory device unreadable... this is from real world NAND memory devices failing from real world use, all of a sudden not wear leveling seems like a suicidal mode of wear... if the entire chip can short out from a single block failing. There's 0 reason for a properly designed flash device to fail like that due to wear, leveling or not. That's just shitty engineering. "Tunnel injection is the quantum tunneling effect, also called Fowler-Nordheim tunnel injection, when charge carriers are injected to an electric conductor through a thin layer of an electric insulator."

you should have said 'IANAEE' for i am not an electrical engineer. because the way NAND ram works it is entirely possible for failure to be a complete and total failure of the device, or at least 512 blocks of the device, if it doesn't create a short that prevents the whole device from working.

first today's flash memory is NAND memory http://en.wikipedia.org/wiki/Flash_memory

NAND memory is written with tunnel injection, which causes charge carriers to be injected into a conductor. http://en.wikipedia.org/wiki/Tunnel_injection

Charge carriers "In semiconductor physics, the traveling vacancies in the valence-band electron population (holes) are treated as charge carriers"http://en.wikipedia.org/wiki/Charge_carrier

So we're using an electric charge, to fill, and create 'electron holes' in a conductor. what could POSSIBLY go wrong, in the real world, rapidly changing if a conductor has electron holes or not, by forcing the electrons in or out of the material ...

so trying to intentionally wear out a NAND memory chip can cause a severe problem whereby instead of creating an electron hole, you've created a short circuit. "A short circuit (sometimes abbreviated to short or s/c) allows a current to flow along a different path from the one intended." http://en.wikipedia.org/wiki/Short_circuit

SSDs will reach price parity on June 15th

[DDumitru]

On June 15th, Mtron will start shipping the 1000 series MLC drives. Put these in an array with the right software and you end up with price/GB parity with 36GB 15K 2.5" SAS drives and about 12x the random IO performance.

HDD Array:

8 Seagate Savvio 2.5" HDDs: $350ea $2,800
    configured raid-10
1 SAS raid controller $600
Total cost for 144 GB $3,400 or $23.61/GB

SSD Array:

6 Mtron 1025-32 2.5" SSDs: $290ea $1,740
    configured raid-5
1 SATA raid controller $250
MFT Software License $1,250
Total Cost for 144 GB $3,240 or $22.50/GB

HDD Performance:
    4K and 8K read IOPS: 250/2000 (single-threaded/multi-threaded)
    4K and 8K write IOPS: 1200

SSD Performance:
    4K read IOPS: 8000/48000 (single-threaded/multi-threaded)
    8K read IOPS: 6000/36000 (single-threaded/multi-threaded)
    4K write IOPS: 40000
    8K write IOPS: 22000

These performance numbers are with the MFT driver in place. Without MFT, the 4K random write performance is about 140 IOPS (>250x slower).

Endurance for these SSDs in this configuration is good enough to overwrite the entire array with random data three times a day (500GB of random updates/day) for about five years.

These drives make a wicked mail server (EasyCo just moved one of it's mail servers mirrored to MLC flash and the difference is amazing).

Sorry for the blatant advert, but SSDs are here now.

Doug Dumitru
EasyCo LLC
http://managedflash.com/
+1 610 237-2000 x2