Slashdot Mirror


DSI Delivers up to 3GB/s with Solid State Disk

olivesaregross writes "'Running at what the company says is 250 times the speed of conventional hard drives, it won't come cheap, but it will be fast. It uses DRAM memory to store data instead of spinning platter hard drives, giving an access time of just 20 microseconds.' It still does use platter-based drives but it's a cool idea anyway. Techworld has another story on it."

214 comments

  1. Pricing by Faust7 · · Score: 3, Interesting

    Traders link to the system over FDDI, T3 or ATM links, and the Eurex back office servers connect via 2Gbit/s Fibre Channel links and switches to the SSDs. The system uses DSI's 3200 solid state disks, with two to eight Fibre Channel ports that can push out 250,000 IOPS - up to 3Gbit/s - and contain 16-64 GB of capacity. There are two hot-swappable power supplies and three hot-swappable drives per DSI 3200. Uptime is five-nines - 99.999 percent.

    Well! A consumer-level version ought to be cheap in about... ten years.

    1. Re:Pricing by Vihai · · Score: 2, Insightful


      I would call it "a hard disk with a big cache"... however... I can have a cheaper disk with a bigger cache just by adding more RAM to my system and it would be much much faster... wonderful eh?

      I may be wrong... but I cannot RTFA.. the site is /.tted...

    2. Re:Pricing by b0r0din · · Score: 2, Interesting

      From the article:

      "Trying to get the general population to think beyond the big RAID systems is our biggest impediment to solid state disk acceptance..."

      That and the hefty pricetag, I'm sure. Obviously the better demand, the better a chance the price will go down, but seeing as RAID stands for Redundant Array of Inexpensive Disks (for some people) may explain why it's so popular. Same reason most people don't drive Ferrari's - it costs like 1000 dollars for an oil change.

      I don't see this having much useful business application for the time being outside of very high enterprise computing, although it's about time. I'm still waiting for optical storage drives.

    3. Re:Pricing by TechnoConfucius · · Score: 1

      I write a trading system that connects to Eurex - and DB performance (determined by disc access and io speed predominantly) is one of the major system issues. This is better than RAM because persistent storage is critical for transactional control and auditing, (imagine the consequences if you are not sure whether a trade has been placed or not!). I've made a careful note to investigate this in detail tomorrow with our infrastructure guys!

    4. Re:Pricing by corbettw · · Score: 4, Funny

      I may be wrong... but I cannot RTFA.. the site is /.tted...

      Apparently, increasing memory access times wasn't the panacea they thought it would be.

      --
      God invented whiskey so the Irish would not rule the world.
    5. Re:Pricing by eht · · Score: 4, Informative

      RAID has a couple of semi accepted meanings,

      The I can mean either Inexpensive or Independent
      and the D can be either Drives, Disks, or Devices.

      However it's always a Redundant Array, which of course makes RAID 0 not RAID, but just a good way to lose even more data faster(as any drive/disk/device failing on RAID 0 takes down all your info).

    6. Re:Pricing by ktulu1115 · · Score: 2, Interesting

      The I can mean either Inexpensive or Independent and the D can be either Drives, Disks, or Devices.

      I never understood that... RAID has historically been set up with SCSI drives, (ATA RAID has been more recently developed/implemented) I would hardly consider them cheap compared to ATA (altough recently I have noticed the price differiental somewhat lessened)

      Regardless, I still use almost all SCSI in my boxes.

      --
      # fuser -v /dev/attention | grep work
      #
    7. Re:Pricing by hyc · · Score: 1

      You could do a consumer version using end-of-life slow DRAMs, but it seems that slow DRAM doesn't stay on the market very long after its primary use (PCs) moves on to faster stuff.

      With UltraATA topping out at 100MB/sec even the cheapest/slowest available DRAM can keep the interface chugging at full speed 100% of the time. A simple 16 or 32 bit bus clocked at 50 MHz (one 16 bit word per cycle since ATA is only 16 bits wide) will be sufficient.

      Throw in a couple batteries (NiMH or LiIon, whatever's cheaper) and you're golden. The amount of power needed to keep the DRAM refreshed is not a big deal...

      The question in my mind is whether it would be good enough to use a solid state drive like this all by itself, or whether it should mirror onto a real disk, or whether the memory should just be used as a hardware cache in front of a completely separate disk drive. I think you still get the most bang for the buck by just using it as a big cache for another disk.

      --
      -- *My* journal is more interesting than *yours*...
    8. Re:Pricing by roblake · · Score: 1

      A bit of history. I invented the solid-state disk concept in 1970 and licensed it to Monolithic Systems, which marketed it as the "EMU" Extended Memory Unit. We expected the performance (on a PDP-11 Unibus) to be 17,000 times faster than the RF/RS-11 fixed head disk it replaced. Based on 17 millisecond rotational latency for the rotating hard drive vs. the 1 microsecond RAM memory. Well, it didn't work that way. Turns out the DEC device driver sucked up a LOT of time and, while performance was markedly increased, it wasn't 17,000 times. Best application was on a submarine, where the rotational latency didn't matter --- but the rotational latency DID! Rob Lake

    9. Re:Pricing by DrSkwid · · Score: 1


      Which is cheaper

      5 x 9 Gb Scsi

      or 1 x 36gb scsi with the same access/throughput speed ?

      --
      There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter
    10. Re:Pricing by Anonymous Coward · · Score: 0

      All of the RAID disk at our shop are Fibre Channel

  2. Not cheap, but fast by Anonymous Coward · · Score: 4, Insightful

    then it must be good. cheap, fast, good, pick any two.

    1. Re:Not cheap, but fast by Anonymous Coward · · Score: 0

      then it must be good. cheap, fast, good, pick any two.

      I'll take "good" and "good" then.

    2. Re:Not cheap, but fast by Anonymous Coward · · Score: 0

      I remember it differently: "Cheap, fast, or big; pick any two."

    3. Re:Not cheap, but fast by Anonymous Coward · · Score: 2, Informative

      Actually the cheap, fast, and good mean different things from what the original poster intended.

      Cheap is cost.
      Fast is time to market.
      Good is reliability/performance/design.

      In other words I can design a syatem that is cheap to make and performs really good but I can't do it fast.

      Or I could do it fast and have it really good but its gonna cost a lot (Think xbox).

      Or I could do it fast, make it cheap but its not gonna be very good. (Think N-gage)

      Finally It can take forever, cost a fortune and be really crappy. (Think Longhorn)

    4. Re:Not cheap, but fast by Erwos · · Score: 2, Insightful

      I believe you mean "cheap, fast, large capacity; pick two". "Good" is not really meaningful.

      -Erwos

      --
      Plausible conjecture should not be misrepresented as proof positive.
    5. Re:Not cheap, but fast by Excen · · Score: 0

      Finally It can take forever, cost a fortune and be really crappy. (Think Longhorn)

      Or Duke Nukem Forever for that matter. . .

      --
      "No beer until you finish your tequila!" -Leela's Dad
    6. Re:Not cheap, but fast by dustmite · · Score: 1

      Doesn't "fast" usually refer to time taken for development and/or to market?

  3. Interesting Uses Possible by Pavan_Gupta · · Score: 4, Interesting

    I can imagine this kind of technology being really applicable in situations where large databases are in use -- but potentially, slightly cheaper then just keeping the entire database in ram. I think it would be interesting to use, but a bit more interesting to play with.

    1. Re:Interesting Uses Possible by Neon+Spiral+Injector · · Score: 4, Interesting

      But the entire DB is in RAM. But it isn't local to the processor, it is on the other side of a FC bus.

      When this makes sense, is when you do want your entire DB in RAM, but don't want to put the required amount of RAM in each node of a cluster. So you just attach each node to this SAN.

    2. Re:Interesting Uses Possible by j.+andrew+rogers · · Score: 1
      "I can imagine this kind of technology being really applicable in situations where large databases are in use -- but potentially, slightly cheaper then just keeping the entire database in ram."

      One of the major problems with databases is how much RAM you can physically put in the machine. You can put terabytes of "almost RAM" on your disk subsystem for a server that maxes out at 16Gb of physical motherboard RAM. If your working set is large such that there is no way you'll be able to practically fit it in RAM on the motherboard, this can serve as an intermediate solution for extending the system RAM rather than bottlenecking on platters.

      It is primarily useful as a pseudo RAM extender, for applications that simply won't fit in the amount of RAM most servers can hold but which need RAM-like access speeds.

    3. Re:Interesting Uses Possible by dAzED1 · · Score: 1
      very, very soon we'll be done with addressable memory as a problem. At least, for a long while.

      and before someone makes a 640k comment, don't. There's a difference. When it was 640k, plenty of people could imagine uses for a gigabyte. How many uses for a petabyte of ram can anyone really justify in the next 10 years?

    4. Re:Interesting Uses Possible by haute_sauce · · Score: 1

      The great thing about this is that you would not have to put the WHOLE database in RAM, just the 'areas' that are used heavily. In Oracle, that would be the TEMP tablespace (where sorting, etc occurs), The REDO areas (self explanitory), and the LOG areas. Just those areas alone (and maybe some indices) would probably increase I/O performance dramatically.

    5. Re:Interesting Uses Possible by j.+andrew+rogers · · Score: 1
      As a point of reference, I could really use about a terabyte of RAM (not that I have it) for research purposes right now and I can easily seeing having a ton of use for a petabyte of RAM because we will eventually outgrow our current needs by a couple orders of magnitude. Access time is a matter of tractability for many algorithm spaces.

      Of course, at this point the real problem starts to become memory latency and bandwidth.

    6. Re:Interesting Uses Possible by Twirlip+of+the+Mists · · Score: 1

      How many uses for a petabyte of ram can anyone really justify in the next 10 years?

      That's not the relevant question. The relevant question is, how many uses for 16,384 petabytes of RAM can you think of?

      (2^64 = 16,384 * 2^50)

      --

      I write in my journal
    7. Re:Interesting Uses Possible by walt-sjc · · Score: 1

      Sure. And when your machine crashes you lose data. Sometimes you need true NV ram disks.

    8. Re:Interesting Uses Possible by haute_sauce · · Score: 1

      Umm-I am assuming you didnt read the original article, that tells of batteries AND HD backup on the DSI disks. Which makes them not only a whole lot faster than HD (alone), but probably a whole lot more reliable.

    9. Re:Interesting Uses Possible by walt-sjc · · Score: 1

      I read the article, but the parent said RAM, not DSI or SSD. A solid-state disk is not RAM even though it uses RAM chips. If you use vague or incorrect terms, nobody knows what you are talking about. There are many other comments for this story talking about normal ram disks - that's what I was thinking when I replied.

      Solid Data's drives (which I have used) do the same thing - nothing new here. Banks of ram chips on a card that make it look like a disk. Internal battery. If power fails, contents of ram are dumped to disk. On power resoration, disk is read back into ram chips.

      SSD's are nothing new. They are expensive, fast, have limited storage, and are good for certain types of applications. Nothing has really changed since they were introduced many years ago other than price, size, and speed (which depends more on the interface rather than anything else...)

    10. Re:Interesting Uses Possible by WIAKywbfatw · · Score: 3, Funny

      See, it's just this sort of comment that has my girlfriend thinking that Slashdot isn't for regular nerds, it's for 10th dan nerds.

      --

      "Accept that some days you are the pigeon, and some days you are the statue." - David Brent, Wernham Hogg
    11. Re:Interesting Uses Possible by addaon · · Score: 1

      Well shoot, I'm still working on my third frank.

      (I'm amazed that I resisted saying 'my second willy.')

      --

      I've had this sig for three days.
    12. Re:Interesting Uses Possible by Ed+Avis · · Score: 1

      Are high-end servers still sold that won't take more than 16Gbyte (I assume you meant 16Gbyte not 16Gb) of RAM?

      You have to compare these expensive RAM disks against two alternatives. Firstly, getting a server with a reasonable number of memory slots and loading it up with X gigabytes of memory; secondly, buying a dozen PCs, sticking four gigs of memory in each and having some wacky Gigabit Ethernet, SCSI, or Fibre Channel attachment to turn each one into a RAM disk. Although I expect that the power consumption for a PC would be higher, and it might take more space.

      --
      -- Ed Avis ed@membled.com
  4. Their website by dot-magnon · · Score: 2, Funny

    Their website or hosting firm isn't very dynamic, at least. Right now, it's not even static. It's slashdotted.

    1. Re:Their website by Anonymous Coward · · Score: 1, Insightful

      We know. But thank God you pointed this out. We'd be completely lost if no one informed us of the Slashdot effect, each and every time it fucking happened.

    2. Re:Their website by Anonymous Coward · · Score: 0

      Didn't survive the slashdotting. Well back to the drawing board.

  5. Uses? by kunudo · · Score: 1, Troll

    What would the use for this be? I can see how it would be nice to have no loading time at all in every game etc, but is anyone except NASA going to use this now? Any ideas?

    1. Re:Uses? by Sch0pehauer · · Score: 1

      It would be disponible to the end user as soon as Internet2 will be. As we have already seen the speed of that network makes work hard to hard disks that try to save the data flux.

      For now we have a 100MB/s speed record across the Atlantic, but it will improve.

    2. Re:Uses? by maharg · · Score: 1

      I have 45Gbytes of *index* (Verity K2) which I'd like to have in RAM, but I only have 8Gbytes of RAM. This device would speed things up a bit I think !

      --

      $ strings FTP.EXE | grep Copyright
      @(#) Copyright (c) 1983 The Regents of the University of California.
  6. Re:Obligatory first post! by JudgeFurious · · Score: 0, Offtopic

    I don't know if I find this to be too offtopic. I mean, technically it IS a possible use if I'm not mistaken.

    I think you got an offtopic because you posted too soon. In another 15 minutes it would have been modded Funny (though it was only marginally funny at best).

    --
    Appended to the end of comments you post. 120 chars.
  7. Looks like those companies... by Anonymous Coward · · Score: 5, Funny

    Are already starting the hard work on getting the components together to meet the minimum specs for running Longhorn. Huzzah!

    1. Re:Looks like those companies... by JudgeFurious · · Score: 1

      Yes they are. While this drive itself wouldn't be capable of running Longhorn I'm sure that what was learned getting here will be put to use when they turn their attention towards the eventual drive they develop that's capable of running it.

      Another decade full of baby steps like this and we're there!

      I know I can't wait.

      --
      Appended to the end of comments you post. 120 chars.
  8. about time... i hate platters by Anonymous Coward · · Score: 0

    ive been waiting for this since hd's were 1st envisioned... platters suck...

  9. Thats great by rberton · · Score: 0

    until the power goes out.

    1. Re:Thats great by Anonymous Coward · · Score: 0

      Wait, are you being clever or stupid?

      As Spinal Tap pointed out, it's a fine line.

    2. Re:Thats great by erikdotla · · Score: 3, Informative

      If you read the PDF brochure, it explains that there is a regular disk inside as well. The device is constantly backing up the DRAM contents to disk, and the device contains battery power which guarantees that in the event of a power outage, there is enough time to fully back up the DRAM contents. So power outages won't hurt it (unless maybe you average more than outages per hour.)

      --
      # Erik
    3. Re:Thats great by hyc · · Score: 1

      The idea seems to be that the size of DRAM and size of disk are the same. If it were my design, I would leave the disks OFF all the time, and have a large battery in place. Whenever the power switch was turned off, or when AC power was cut, then I'd spin up the disk and flush the RAM out.

      There's no reason to have the disks spun up eating electricity and generating heat and noise in normal operation. With today's low-voltage parts, a couple of LiIon batteries could probably keep the entire memory array live for so long you'd never need to flush to disk anyway.

      --
      -- *My* journal is more interesting than *yours*...
  10. Re:Obligatory first post! by Neil+Blender · · Score: 2, Funny

    Ultra fast access prOn!

    It will be great for you since finally the picture will load before you shoot your load.

  11. Is just me... by amarcuss · · Score: 1

    ...or does this blurb makes no sense and contradict itself?

    1. Re:Is just me... by Anonymous Coward · · Score: 0

      "It uses DRAM memory to store data instead of spinning platter hard drives, giving an access time of just 20 microseconds.' It still does use platter-based drives"

      Im confused ;(

    2. Re:Is just me... by Anonymous Coward · · Score: 1, Informative

      Backups. When the power goes away, the data in DRAM will, too. Best to store it somewhere, first....

    3. Re:Is just me... by Anonymous Coward · · Score: 0

      Then it should say it uses DRAM in *parallel* with platter based drives. It takes about 5 seconds to realize that the way it's written is confusing and about another 5 seconds to reword it more clearly.

    4. Re:Is just me... by Glonoinha · · Score: 2, Informative

      The way I understand it the system shadows the entire hard drive in DRAM. Reads come from the DRAM, writes first update the DRAM and are written back to the drive.

      Envision an 8G SCSI hard drive with 8G of cache, pre-populating the cache when you turn on the computer. Same idea.

      In fact this is a hardware manifestation of the Superspeed software that used to be marketed by Cenatek (not sure if they still do - Google it.)

      --
      Glonoinha the MebiByte Slayer
  12. Obligatory comment... by beavis88 · · Score: 3, Funny

    Maybe they should be serving their website from SSDs

    *waits for the groans*

    1. Re:Obligatory comment... by antic · · Score: 1

      Has anyone ever wondered about the effect that is a "slashdotting"? See, it's fact that no one will ever "RTFA". And it's also a foregone conclusion that anything mentioned on /. is overwhelmed by traffic. So what's happening in the middle there?

      And back (slightly) on track -- with the technology arriving yearly with faster and smaller hard-drives (mostly used in digital music devices) why do people still have giant (even a midi case is massive) cases on their desktops? Is it only due to standardised components, sizes, plugs, etc?

      Because working on the inside of a PC is ugly stuff. IDE cables are bastards. Sharp edges and tiny switches are everywhere. Now, not everyone needs to get in there to tinker, but surely there are enough geeks around the place to warrant more usable hardware? I mean, someone's out there making money from selling glowing cables to people fancifying their computers, and when Slashdot isn't showing endless ads for Microsoft products (or taunting everyone with ads saying Windows is faster than Linux) it's often advertising modifications to make computer cases look like plastic fish tanks...

      --
      'Thats they exact same thing a banana wrench monkey.'
    2. Re:Obligatory comment... by Anonymous Coward · · Score: 0
      There are two types of slashdot readers... Those who read the articles (the vast majority), and those who comment on them (a tiny minority). Unfortunately there isn't nearly as much overlap between the two as you'd hope, hence the apparent contradiction wrt overloaded servers and uninformed comments.

      It's even worse because the slashdot moderation system heavily rewards early comments, so there is more incentive for those who base their self-worth on their comments' moderations to not bother reading the article, or in some cases to not even bother reading the entire slashdot posting...

    3. Re:Obligatory comment... by raodin · · Score: 1

      I don't know what kind of computer you're using, but mine doesn't really have any sharp edges or tiny switches, and has just a few jumpers (Optical drives, clear CMOS, possibly a couple others that I never touched or cared about). And no, it doesn't look like a plastic fishtank, and certainly not a glowing one. :)

    4. Re:Obligatory comment... by Blastrogath · · Score: 1

      And back (slightly) on track -- with the technology arriving yearly with faster and smaller hard-drives (mostly used in digital music devices) why do people still have giant (even a midi case is massive) cases on their desktops? Is it only due to standardised components, sizes, plugs, etc?

      Minitureized stuff costs more. The mini hard drives that store 5GB or so cost about as much as a regular size drives 40X their capacity.

      Digital music devices are made to be portable, and PCs are not. A better comparison would be digital music devices vs. laptops/tablet PCs.

      That being said, I think a lot of the difference is because of PC's being expandible, and that a PC is much more capable than a MP3 player.

      --
      "The price good men pay for indifference to public affairs is to be ruled by evil men." -Plato
    5. Re:Obligatory comment... by Anonymous Coward · · Score: 0
      So what's happening in the middle there?
      2. ??? Of course. ;)
    6. Re:Obligatory comment... by DoraLives · · Score: 1
      why do people still have giant (even a midi case is massive) cases on their desktops?

      Some of us prefer a nice roomy engine compartment as opposed to something else. It's always more fun to work someplace where there's room for not only the item you're working on, but also your hands and the tools they're holding.

      --
      Is it fascism yet?
    7. Re:Obligatory comment... by cujo_1111 · · Score: 1

      I wish I had mod points...

      --
      If I point out that you are incorrect, making me a foe does not make you any more correct.
  13. A new requirement... by D-Cypell · · Score: 3, Funny

    I wonder how long it will be before microsoft add this technology to their list of requirements for longhorn!

    1. Re:A new requirement... by Anonymous Coward · · Score: 0

      On Anarchist Slashdot, that joke makes itself.

  14. SSD is a niche technology at best by Anonymous Coward · · Score: 4, Informative

    Let's see:

    Imperial folded, Platypus folded, Solid Data is barely hanging on and Texas Memory survives on defense contracts.

    SSD is a great technology, yes.
    SSD makes commercial sense, no.

    How many more VCs can be fooled into investing into SSD startups?

  15. Why is this significant? by John+Jorsett · · Score: 3, Insightful

    We pretty much expect things to get increasingly bigger and faster, so is another RAM-based pseudo-disk that big a deal?

  16. 26.6666666 times faster by Anonymous Coward · · Score: 2, Interesting

    Wasn't there already a solid state transfer rate of 80GB/s reported from SGI/Cray a year or so ago?

  17. This will be great! by Kjuib · · Score: 0

    I can trade in my 2 ata Raid drives and just get one of these "Mothers" Never have to wait for load, and just rip all my game CDs to my harddrive and no mo downtime... I will name my computer lighting 0 for it is fast!

    --
    - Your stupidity got you into this mess, why can't it get you out? -Will Rogers
  18. Re:Not cheap, but fast.. my pick.. by cbreaker · · Score: 1

    Cheap and Fast.

    --
    - It's not the Macs I hate. It's Digg users. -
  19. funny by Anonymous Coward · · Score: 1, Funny

    The TechWorld article (titled: Welcome the Ferrari of disk drives) has an ad for the Iomega Rev on the page which might confuse the average schmoe out there that it's the Iomega that is the Ferarri, shudder.

    (click-death!)

  20. RELIABILITY!!! by NineNine · · Score: 4, Insightful

    Screw speed. At least for me, that's not an issue. I want a r-e-l-i-a-b-l-e hard drive. I've tried all the brands, but they all come down to this: You have moving parts. It's going to break, eventually. The bearings will go. The head will hit a platter, etc. Personally, I've been saying for years that a solid state hard drive will be the next big boost in PC technology, and I think this is the beginning.

    I don't understand why the company isn't touting reliability. If I have a slow hard drive... eh. No biggie. I wait an extra second. If I have a hard drive crash, that's potentially days of lost work and business, even more if a backup failed recently. I'll be buying these just as soon as I can afford them. With these drives in place, the next reliaibility bottleneck are the stupid little cooling fans failing. Electronics (printed circuit boards, chips) rarely fail on their own. It's almost something with moving parts (like a fan) that leads to their death.

    To me, this is the most exciting advance in computing since Ethernet.

    1. Re:RELIABILITY!!! by Lifewish · · Score: 1

      As a home computer user, I've never yet had a hard drive crash. About how paranoid do my fellow /.ers feel I should be about this?

      --
      For the love of God, please learn to spell "ridiculous"!!!
    2. Re:RELIABILITY!!! by Jetboy01 · · Score: 1

      I'm sorry, Lifewish cant reply right now, hes busy restoring from backups after his hard drive crashed just seconds after posting the last message

    3. Re:RELIABILITY!!! by Anonymous Coward · · Score: 0

      A major factor is speed with large databases. I recall a /. post about the US government buying a 2.5 TB RAM disk to speed up searches of massive databases for threat information. http://slashdot.org/article.pl?sid=04/03/09/195223 4&mode=thread&tid=103&tid=137&tid=164&tid=185&tid= 99

    4. Re:RELIABILITY!!! by Anonymous Coward · · Score: 0

      It's likely you replace your entire computer before the hard drive reaches its failure point--many home users do.

      However if you're a build-your-own-computer type or just a cheapskate, those drives move from machine to machine as long as they work--and you get more years out of your hardware, but you also run a greater risk of having a drive go bad.

      Nothing with moving parts (hard drive, fans, power supplies) should be expected to last longer than five years. Throw CRTs in the mix too.

      I'm not saying they can't last longer (I once had a CRT last seven years on me with no problems), but the failure rate increases to quite a high level at that point.

    5. Re:RELIABILITY!!! by neurojab · · Score: 4, Insightful

      >I don't understand why the company isn't touting reliability.

      Traditionally, DRAM-based storage units are LESS reliable than hard disks. Why? Power loss. Yes, you can always create a massive UPS, but to be really considered "stable storage", you need to be able to store data without power for years on end. The advantage here is that the unit writes data to its hard disks, giving you some assurance that you won't lose your data even then. That turns the RAM in the unit into a giant cache. This helps with read operations, but does nothing for writes. Besides, if the hard drives fail in this unit, the unit still fails. That makes it no more reliable than RAID.

      >To me, this is the most exciting advance in computing since Ethernet.

      This concept has been around since before Ethernet. The concept of storage in solid-state isn't new. Even using RAM for a hard disk isn't new. Ever run VDISK in an expanded-memory DOS system? This concept was available in the higher-end comptuer world far before that.

    6. Re:RELIABILITY!!! by Anonymous Coward · · Score: 0

      Insightfull my ass.

      Your talking about a system based around dram. If your power goes so does your data. Dram is also screwed by alpha particles. If you want reliability look into server class disks or mirroring. Either one is cheaper and more reliable then a ram disk.

      The only thing a ram disk has going for it is speed.

    7. Re:RELIABILITY!!! by Anonymous Coward · · Score: 0

      RAID offers you the reliability - a failure of a single unit does not compromise the performance of the whole. Multi plexes/sub mirrors in straight RAID 1 would add the resiliance you want. RAID 1+0 adds a bit of speed to the equation.
      With a standard RAID 1 configuration - 2 sub mirrors - I've not seen a single need to restore from backup from over 300 machines and 3000+ drives in the last year (since I've had exposure to this environment). We get about 3 drive failures a week.
      If your data is that important to you splash out a few extra dollars on a extra drive, and, if you want the availability/resiliancy/speed, an extra controller. You're looking at, what, $80 for the drive, $160 if you want a spare sat around for when the disk dies, $40 for the controller? Software RAID 1 comes as standard on Linux, Windows 2K,XP Pro, 2K3 and Solaris. There really is no reason why you shouldn't be using a resiliant storage solution for valuable data, even on a budget.

    8. Re:RELIABILITY!!! by Bilestoad · · Score: 1

      As a home computer user I had never had a hard drive crash despite having used old drives, second hand drives and the same drives without any special cooling for years on end, beginning in the days when a 286-12 was something special. Then one day last year two drives (both less than 6 months old, made by different manufacturers, one on IDE and one in a firewire enclosure) both died. I lost important stuff that can't ever be replaced.

      Don't be too confident, it will happen to you one day. If you don't have everything important on some other media stored in another location you're asking for trouble.

    9. Re:RELIABILITY!!! by menscher · · Score: 1
      I don't understand why the company isn't touting reliability.

      Probably because they are? Their specs say 99.999% uptime all over them.

      Too bad the storage is so teeny. Only 8-64 gig? What are we going to fit on that? Oh well, I guess it has applications for databases. But certainly nothing where you could survive waiting for your raid array to fetch the data.

    10. Re:RELIABILITY!!! by .pentai. · · Score: 1

      Actually, it does help with writes too, because the "drive" is the ram, and the harddrive it writes to is the backup...meaning everythign is used from ram, and the ram is paged out to disk at intervals to keep a consistent backup. These things generally also have batteries which supply enough juice that should power go out they can write out the rest of ram to the drive and shutdown semi-cleanly...

    11. Re:RELIABILITY!!! by Deraj+DeZine · · Score: 1

      I agree wholeheartedly; bring on the hard drives with r-e-l ... y ... ability. My life has been going down hill ever since I got cut from the cheer squad...

      --
      True story.
    12. Re:RELIABILITY!!! by pr0c · · Score: 1

      As a home computer user, I've never yet had a hard drive crash. About how paranoid do my fellow /.ers feel I should be about this?

      Quickest way to do that without opening it up and dumping sand in there is to defrag daily... there are some idiots in the office I have told over and over not to do that (BUT IT SPEEDS IT UP! heh..) and one of them just lost their entire harddrive. I've resorted to renaming the defrag utility now....

    13. Re:RELIABILITY!!! by Atmchicago · · Score: 1

      To me, this is the most exciting advance in computing since Ethernet.

      First you need ethernet to get your pr0n from the internet, and then you need the solid state hard disks to access it fast! And you won't even need to back it up!

      --

      You can lead a horse to water, but you can't make it dissolve.

    14. Re:RELIABILITY!!! by IdleTime · · Score: 1

      If you need anything larger, I suggest you fork over a few millions for one of these babies. Just 2.5Tb of solid state memory disk and with a pricetag of $4.7 million ....

      --
      If you mod me down, I *will* introduce you to my sister!
    15. Re:RELIABILITY!!! by GlassHeart · · Score: 1
      I don't understand why the company isn't touting reliability.

      You're in the minority. Most people's data are not worth the expense of dramatically more reliable hardware. In fact, they apparently aren't even worth the trouble to making regular back-ups for.

      If I have a slow hard drive... eh. No biggie. I wait an extra second.

      No, a slow hard drive can preclude entire applications, such as video capturing.

      I'll be buying these just as soon as I can afford them.

      And that's why not many people are selling it.

    16. Re:RELIABILITY!!! by AbRASiON · · Score: 1

      If the users are needing to run defrag because it's slow - perhaps you're not giving them good enough boxes.

      Plus a defragged disk once a month will SAVE wear and tear from extra head movement of fragmented files.

      You sound like one of the HP monkeys who support us (external support contract)

      Short sighted - any sign of a user with half a brain - lock his PC down so he can barely right click

      - Bravo......

    17. Re:RELIABILITY!!! by juju2112 · · Score: 1
      He said they defragged daily, not monthly.

      I assume the implication is that doing it so ofen wears on the drive too much. And really, doing it so often is really pretty pointless. The data doesn't get mixed up that fast. It'd be like compulsively throwing your cd collection on the floor and re-alphabetizing it every day.

    18. Re:RELIABILITY!!! by Slashamatic · · Score: 1
      Drives may rotate but generally speaking the high-end is better than memory because you can't pull the plug on memory. What we have here is a solid-state ram backed by an HD combo. The unit can still fail so generally they are doubled up. As someone else has commented here, the Fibre Channel may be new but the tech is about 15 years old.

      The main problem with drives at the moment is insufficient cooling. Sprin them fast and drive the heads hard and they will get hot. Commercial units don't tend to be that different, but at least they ensure that they run cooler.

    19. Re:RELIABILITY!!! by warrax_666 · · Score: 1

      Since they're battery backed, you could also just have a RAID of SSDs. That should lower the likelihood of failure to tolerable (i.e. "not gonna happen") levels.

      --
      HAND.
    20. Re:RELIABILITY!!! by boots@work · · Score: 1

      Even error-correcting memory may suffer an uncorrected failure in the multi-year lifespan of a disk. I don't know if RAM would be more reliable byte-for-byte than platters. In fact, I think either one has an MTBF on the order of a few years for typical sizes, and disks are a lot larger.

      Disks are so much cheaper than you can simply have mirrors and backups. Buy ten disks, and put them in mirrorred pairs. Use something like rdiff-backup to make backups and you can likely keep every version of your filesystem for the last year.

      You need to protect against user error and software failure, not just hardware failure. Accidental deletion is far more common than drive failure, and only some kind of backup or snapshot system will protect you against that.

  21. Google Cache by DeathToBill · · Score: 4, Informative

    Hmmm, one of the fastest slashdottings in recent history, methinks. Google cache is here.

    --
    Slashdot - News for Nerds, Stuff that Matters, in ISO-8859-1 Has just realised that beta makes this signature redundant
  22. Really just a disk with a huge cache.. by whoever57 · · Score: 1

    These units store the data in RAM and on disk (a RAID array). At some point the data has to be written out to disk and this will not be as fast. Alternatively, it's possible that when powering off, the entire contents fo the cache could be written to disk -- the time to power off would be quite large though. I see at least one unit has it's own UPS to manage this.

    --
    The real "Libtards" are the Libertarians!
    1. Re:Really just a disk with a huge cache.. by Anonymous Coward · · Score: 2, Informative

      A UPS is one way to keep the data going, but it's much simpler and more effective to use the method that Sun used years ago with its Prestoserve disk-cache boards years ago: you put a little lithium watch battery on the board with the RAM. If the system loses power, the battery powers the RAM for, I don't know, a year or two. Then you also build in a little software-accessible battery monitor that tells you when it's time to replace the watch batteries.

      Of course, you have to have some software so that after a crash or power failure this disk cache can be properly synchronized with the real disk. But that's not super difficult.

      (Also, if anyone wants to know, the purpose of the Prestoserver was to accelerate NFS performance. Especially on the old NFS v2, write performance was bad because the protocol specified the data must be committed to nonvolatile storage before the NFS request could be acknowledged. So you have a lot of threads sitting around that could be serving other NFS requests but instead they're waiting on something to be written to disk. The solution was Prestoserve, basically a queue of things to commit to disk but stored in battery-backed RAM. It was goofy, but it provided all the benefits of write cache with basically none of the risk, so that was nice.)

    2. Re:Really just a disk with a huge cache.. by stephenisu · · Score: 1

      the time to power off would be quite large though.

      Well given the operating hours of this particular application (stock market), they only run 12 hours a day anyways.

      --
      Sigs? We don't need no stinking sigs!
    3. Re:Really just a disk with a huge cache.. by bofkentucky · · Score: 1

      Thanks, I noticed the prestoServe when I was looking for parts for an Ultra2 that is soon to become my testbed machine at work, got distracted and didn't google for it. I'm just going to be glad to see that SparcStation 5 disappear from my office, it is just a little underpowered for my tastes and very little disk space.

      --
      09f911029d74e35bd84156c5635688c0
  23. I know it's not quite the same as a ramdisk but by foidulus · · Score: 2, Informative

    This idea seems to have been around for a while. I remember seeing a few years ago a hd controller that you could plug standard ram into to act as a fast cache. Now granted this is on a much larger scale, but.
    It is still cool though :P

  24. Slashvertisement? by stratjakt · · Score: 4, Interesting

    It still does use platter-based drives but it's a cool idea anyway.

    From the article, I gather these are merely SAN boxes with up to 64GB of DRAM, fiber channel output, and 3 hot-swappable hard drives that act as backup.

    Has a record been broken? Has anything special happened? Sure this is high-end stuff, but it doesnt seem new or particularly exciting.

    --
    I don't need no instructions to know how to rock!!!!
    1. Re:Slashvertisement? by LoocSiMit · · Score: 1
      From the article, I gather these are merely SAN boxes with up to 64GB of DRAM, fiber channel output, and 3 hot-swappable hard drives that act as backup.

      Doesn't that give you the horn?

      --
      Intellectual Property
      Intellectual: of the mind
      Property: that over which one has control
  25. Noise by zepmaid · · Score: 0

    What about the HDD noise?? Are solid state disks quieter??

    1. Re:Noise by Anonymous Coward · · Score: 0

      Solid State: A device which is able to control current without the use of moving parts.

      It would be safe to assume that, barring fan noise, it will be considerably quieter.

    2. Re:Noise by Erpo · · Score: 1
      What about the HDD noise?? Are solid state disks quieter??

      From the link:
      The primary storage media is the SDRAM, with three independent disk drives providing secondary storage [...] Two redundant and hot swappable power supplies, fans and multiple drives allow the 3200 to survive multiple points of failure

      Solid state components don't make any noise because there are no moving parts. This device incorporates solid state storage along with fans and normal hdds, so it's not silent.
  26. Re:Obligatory first post! by JudgeFurious · · Score: 1

    And..........yes. Funny. Modding patterns I believe are as predictable as anything ever gets

    --
    Appended to the end of comments you post. 120 chars.
  27. dram by ejaw5 · · Score: 2, Funny

    I certainly hope they're using RAMBUS memory..unless they want to be sued for "not wanting to pay loyalty fees"

    --

    $cat /dev/random > Sig
  28. Logical Uses by Wasteofspace · · Score: 3, Funny

    This technology would be purely for database read speed at the present moment, though down the track I would imaging that this technology will be used to rapid access, Read only, operating system storage. Basically to make idiot proof operating systems for future, high speed computer systems.

    "I have a problem with my computer, it says boot disk invalid"

    "Please remove the cartridge from the front of the PC and replace it with one that will arrive at your door shortly, and don't worry, you wont lose any information"

    1. Re:Logical Uses by prog99 · · Score: 1

      Indeed but with properly configured buffer pools (or whatever your dbms provider cares to call them) anything "hot" should be locally cached anyway.

      Just hope your developers dont decide to issue a few too many tablescans.

  29. Caching? by amalcon · · Score: 2, Interesting

    Is this sort of thing more or less expensive than plain ol' RAM? If it costs more, then just caching 3gb of data from disk into memory at bootup is more cost-effective. If it's cheaper, then perhaps people will start using this technology for swap space, etc. In any case, I've been waiting to see an HDD using solid-state RAM for quite some time now. If we're lucky, it'll be cost-effective before too long.

    --
    -Amalcon
    1. Re:Caching? by morcheeba · · Score: 1

      This device can get away with using much slower RAM than your desktop, so the bulk price should be cheaper. Also, it can use many many banks (because it is custom-designed), while your desktop has a limited number of slots (and a limited number of slots may forec you to use more expensive 1GB DIMMS rather than what's the best value).

      If it weren't for those two significant factors, then regular-old-ram would be a better investment. But then, throw in the fact that this is sharable between many machines on a FC loop, and it does something better than normal CPU RAM could do.

    2. Re:Caching? by raodin · · Score: 1

      I doubt it. I'm pretty sure I saw something like you're describing years and years ago (think 486/68040 age) and it still isn't cheap...

  30. stop bitching about disk cache by Anonymous Coward · · Score: 0


    All the people who complain that Linux 2.6 uses too much swap for disk cache can buy one of these for a measly cost, set swappiness to zero, and shut the hell up.

    1. Re:stop bitching about disk cache by Anonymous Coward · · Score: 0

      If the situation were flipped and it was Windows instead of Linux, you wouldn't be making justifications that required the purchase of a fucking solid-state hard drive--you'd just be bashing "M$." Nice.

  31. Obviously it's not that fast by GoClick · · Score: 3, Funny

    Obiviously it's not that fast, it's already /.ed

    A number of other companies have been making DRAM disks for several years

  32. Whats new about this? by Kenja · · Score: 3, Informative

    Solid state drives have been around for a long time. Hell, the old RocketDrives could hit 4GB with four 1GB RDRAM sticks.

    --

    "Have you ever thought about just turning off the TV, sitting down with your kids, and hitting them?"
    1. Re:Whats new about this? by radish · · Score: 1

      Or even more to the point, how is this different from the solutions people like EMC have been selling to the enterprise for years now? The idea of tiered storage (ram/disk/tape) with a unified interface is not at all new.

      --

      ---- Den ene knappen er powerknapp, den andre er Bender voice knapp "Bite My Shiny Metal Ass"

  33. barebones ramdrive by Incy · · Score: 2, Insightful

    You think by now. Someone would have started selling these solid state ramdrives with empty RAM sockets & simply support cheap off the shelf SDRAM. No firmware limit of 4gigs or anything. Just a bunch of SDRAM slots on the logic to address them. Is the hardware *that* expensive?

  34. There's hope by GoClick · · Score: 1

    There's hope that with enough public ridicule these problems would get fixed... Also it looks like he posted pretty quick so he didn't have time to read the comments you boob

  35. You asked for it. by Anonymous Coward · · Score: 0

    *groan*

  36. Not a solid state disk. by Anonymous Coward · · Score: 0

    This is NOT a solid state disk. It's a SAN server with a cache as big as the disk array (which is fairly small).

  37. Actually by Anonymous Coward · · Score: 0

    This is chump speed when you start talking about true high end computing, the kind most slashdotters have never heard of. If you think DDR is fast memory or gigabit would make a good "interconnect" for a supercomputer or that a supercomputer even uses TCP/IP at any point (ie rather than a host controller interfacing via another method), you're not playing in the same ball park. top500.org is a joke

    1. Re:Actually by Anonymous Coward · · Score: 0

      What the hell are you talking about?

  38. DSI Delivers up to 3GB/s... by ForestGrump · · Score: 4, Funny

    damn and I thought I could get a 3GB/s DSL line...

    --
    Is it true that more people vote for the winner of American Idol, than vote for the president? -Ali G.
  39. Re:Obligatory first post! by morcheeba · · Score: 4, Funny

    At 3 GB/sec, you can get 3000x3000 uncompressed video at 100 Hz refresh rate, with 1600 channels of uncompressed 96kHz 32-bit sound. Or compress it and get 35000x35000 resolution.

    Pure bliss. But, at that rate, the largest drive (64GB) would only last 21 seconds.

  40. Consumer edition by swb · · Score: 5, Interesting

    How many 1G sticks of RAM could you put into your standard Firewire/USB2 enclosure? Why couldn't someone make a USB2/Firewire/SCSI enclosure that the host system saw as a mass storage device but was actually just a smaller version of the above? It might be really useful for some DB applications, video editing, etc.

    I can't imagine that an enclosure of that type would run more than $500, plus the cost of the RAM that went into it. It might not be consumer cost effective, but it could be worthwhile at the prosumer or low end, where the RAM disks shown on /. are almost never affordable but by the richest organizations.

    1. Re:Consumer edition by Anonymous Coward · · Score: 1, Interesting

      It ought to be significantly less than $500, closer to $300. Consider that you all you need is:

      1) A nimh battery for backup
      2) Sockets for 16 dimms
      3) Physical port for firewire
      4) FPGA to talk firewire out one end and ddr sdram out the other and battery recharge/discharge control out the third end.
      5) Board and case.

      Economy of scale could push something like this down to well under $50 without RAM.

    2. Re:Consumer edition by CylanR77 · · Score: 2, Insightful

      What's the point?

      Any speed gain you'd get from using solid state media would be lost if the device is hooked up to such a bus. And if you really want something small, think about the iPod. It holds many gigabytes of data in a handheld form factor; all it has is a tiny hard drive.

      Also, RAM needs a constant supply of power to keep it useful. If the device is going to be portable at all [and why shouldn't it be, if you hook it up to Firewire or something similar?], it's going to need batteries, something that a hard disk based device wouldn't need to be encumbered with.

      --
      http://cylan.deviantart.com/gallery/
    3. Re:Consumer edition by Anonymous Coward · · Score: 0

      What would be the point in building a really fast storage device and attaching it with a slow interface?

  41. Article Text by Anonymous Coward · · Score: 1, Informative

    DSI's Solid State Disk products provide the easiest and fastest way of removing I/O bottlenecks and increasing system performance. Utilizing a DRAM architecture for storage rather than traditional magnetic disks, these high-performance devices are capable of delivering data of up to 3 GB/s! With no latencies associated with head positioning or seek times, the resulting read/write operations are hundreds of times faster than conventional rotating disks.

    DSI's current line of Solid State Disk products include the DSI3200, and the DSI2100. With no changes to the existing operating system or processor hardware, installation is easy and usually performed in less than one hour!

    Accelerating the Application

    Computer systems and networks have increased exponentially in both speed and performance throughout the 'Digital Age'. Conversely, storage device speed-- measured by seek time, operations per second, total bandwidth, and other mechanics--have remained relatively stagnant.

    These storage device limitations have created a significant bottleneck, resulting in a substantial performance gap over the years. In high-demand networks, this performance gap is painfully obvious--the fastest processors in the world are underutilized if the storage device can't carry out its orders fast enough. And in situations where "hot-files" and databases are constantly read from or written to by multiple sources across a network, this bottleneck creates a crisis for the entire network infrastructure.

    Bringing storage devices into the 'Digital Age', solid state disk technology removes the I/O bottleneck by replacing hard disk drives with high speed circuitry. Instead of a rotating disk, a solid state disk uses memory chips (typically SDRAM) to read and write data, resulting in full utilization of existing processors and bandwidth. With an access time of only 20 microseconds, SDRAM delivers data 250 times faster than conventional disk drives. Uncompromised SDRAM data integrity is maintained through both battery backup and redundant disk drives, so your data is always protected.

    The 3200 has been designed to achieve a 99.999% expected up time, and includes redundant data storage media to ensure data integrity. The primary storage media is the SDRAM, with three independent disk drives providing secondary storage as well as optional data mirroring (active backup). Writes are performed to all four media while read operations are performed by the SDRAM (active backup option), thus solving the SDRAM volatility issue--three disk drives each have an updated copy of all data sets even when the 3200 powers off. Two redundant and hot swappable power supplies, fans and multiple drives allow the 3200 to survive multiple points of failure while maintaining data integrity. Additionally, 3 internal batteries are used in the event of a power failure. By utilizing a 3 GB/s bandwidth and 250,000 I/Os per second, the 3200 gives you the best of both worlds; fast data access and high data reliability.

    DSI3200 BENEFITS:

    Accelerate Database Applications
    Eliminate I/O Bottlenecks
    Improve Response Times
    Data Mirroring
    Open System Support
    "Plug-and-play " Functionality
    99.999% Performance Reliability
    SDRAM with Built-in ECC Circuitry

    DSI3200 SPECIFICATIONS:

    Sustained IOPS: 250,000
    Capacity: 8 GB to 64 GB
    Bandwidth: 3 GB/s
    Access Time: Less than 20 sec Latency
    Size: 5.25"(3U) x 25"
    2 Hot-Swappable Power Supplies
    3 Redundant 30-minute Internal Batteries
    Fibre Channels: 2 GB (2 to 8 ports)

    DSI3200 brochure
    Testimonial: Mid-State Bank
    Testimonial: Texas State Bank
    Testimonial: TrustCompany Bank
    Testimonial: Bar Harbor Banking
    (pdf, 250 kb)
    (pdf, 132kb)
    (pdf, 196kb)
    (pdf, 176kb)
    (pdf, 71kb) Interactive Web Management Demo (Flash, 90kb)

    The 2100 was designed for today's Storage Area Network (SAN) environments which is demonstrated by its ability to support redundant fabric switch connections for high fabric a

  42. Used interface? by Karamchand · · Score: 4, Insightful

    What interface are they using? Even the fastest SCSI can't provide 3GB/s!

    1. Re:Used interface? by chabotc · · Score: 1

      The limitation on SCSI bandwidth is not the chipset or the PCI bus it's connected to, but the 20/40/80/160mbs it can send over the cable it connects it's drives with.

      There's no reason what so ever why you would use a 'scsi' cable to hook up the memory to the chipset.. So limitations you assumed would apply, do not apply

    2. Re:Used interface? by chabotc · · Score: 4, Informative

      Oh and reading in the article it actually tells you whats used to obtain the 3gb/sec

      "with two to eight Fibre Channel ports that can push out 250,000 IOPS - up to 3Gbit/s"

    3. Re:Used interface? by jonbrewer · · Score: 1

      What interface are they using? Even the fastest SCSI can't provide 3GB/s!

      FC-AL, aka Fibre Channel:

      http://www.enterprisestorageforum.com/technology/n ews/print.php/3295601

    4. Re:Used interface? by Donny+Smith · · Score: 1

      >What interface are they using? Even the fastest SCSI can't provide 3GB/s!

      They're using FC it said they've got 2 to 8 ports on one of their boxes, which makes it more than enough (one server, though, couldn't use up all this bandwidth).

  43. Price? by Xaroth · · Score: 1

    The site's /.'ed, and probably not listed in plain sight anyway. How much are we talking for one of these babies?

  44. Re:Obligatory first post! by LilGuy · · Score: 2, Funny

    That's about 10 seconds more than anyone should need.

    --

    You're nothing; like me.
  45. Now that's a drive... by theendlessnow · · Score: 4, Funny

    ...for Longhorn!!

  46. Re:Obligatory first post! by CowboyShit · · Score: 0

    So basically what you're saying is that this is the pr0n device that most of Slashdot has been waiting for all their lives?

  47. Solid State Discs have been around for a while by Burdell · · Score: 4, Interesting
    I've got a mail server that uses an SSD for the mail queue filesystem. It is great for that because the random I/O transactions per second rating is 10,000 (vs. a typical hard drive thrashing hard at 150 tps).

    The SSD we have is a Nitro!Xe from Curtis, Inc.. It looks like a standard 3.5" wide 1" high Ultra2 SCSI drive with an 80 pin SCA connector. We have a 2G model with a 2.5" notebook drive for backup (it has a battery to dump RAM to disk on power off) and it greatly improved the performance of our mail server (high performance mail queue is all about I/O TPS).

    1. Re:Solid State Discs have been around for a while by Anonymous Coward · · Score: 0

      And how much does this thing cost? Are there any resellers with a price list out there?

    2. Re:Solid State Discs have been around for a while by idiot900 · · Score: 1

      I've got a mail server that uses an SSD for the mail queue filesystem. It is great for that because the random I/O transactions per second rating is 10,000 (vs. a typical hard drive thrashing hard at 150 tps).

      Can I ask why a 2GB SSD was better than buying another 2GB of RAM for your mailserver, and putting your mail queue on a ramdisk?

    3. Re:Solid State Discs have been around for a while by vadim_t · · Score: 2, Informative

      Probably because smart mail servers will always ensure that the file has been written to disk before confirming delivery. In these conditions, gigabytes of cache won't help at all, the mail server's still waiting for the disk.

      The advantage of these systems would be that they're insanely fast, and have a battery in case power is lost. Since they emulate a hard disk, the mail server thinks everything has been safely written, and since there is a battery this works pretty well.

    4. Re:Solid State Discs have been around for a while by Burdell · · Score: 1
      Mail is important to people now, and you don't want to just lose a message. The way most mail servers work is that at the end of the SMTP DATA phase, before returning a successful response code, the server flushes the mail queue file(s) for that message to disk (i.e. they don't just close the files; they tell the OS to flush the blocks for those files to disk and don't return until they are written). That way, in case of crash, power failure, etc., the message is not lost. Then the server tells the other end "I've got it" by returning a success code.

      If it were just in the kernel disk buffers or a RAMdisk, then a system crash or power failure would cause that message to be lost completely. Customers don't like lost mail (sure, crashes and power failures don't happen often, but do you want to explain that to the customer that didn't get the job offer, picture of his grandkids, etc.?).

    5. Re:Solid State Discs have been around for a while by silas_moeckel · · Score: 1

      Because per the SMTP RFC(working out of my head here in case anybody wants to find the URL) you are supposed to write the mail all the way out to disk before you accept it. That way even if the power failed at that instant you would have the mail sitting snug on the disk. The worst thing that should happen is a duplicate is sent.

      --
      No sir I dont like it.
  48. Yes, But These Guys Have More Capacity by osewa77 · · Score: 1
    From a linked article:
    Texas Memory System's OEM, Dynamic Solutions International, which has over 18 years of experience delivering solid state disk solutions to the enterprise, installed a 2.5 Terabyte Tera-RamSan at a customer site to accelerate critical database applications, metadata and to maintain a technological edge.
    Texas Memory Touts Largest SDD Installation
    The system uses DSI's 3200 solid state disks, with two to eight Fibre Channel ports that can push out 250,000 IOPS - up to 3Gbit/s - and contain 16-64 GB of capacity. There are two hot-swappable power supplies and three hot-swappable drives per DSI 3200. Uptime is five-nines - 99.999 percent.
    (from the Techworld article)
  49. "Slashvertisement?" by Anonymous Coward · · Score: 0

    Man, we've been seeing more than a few of these here lately. Looks like you can fill in the blanks during a slow news day nicely with some commercial offerings and please the offerers at the same time. The /. editors have probably received some cool new toys recently...

  50. Wafer-scale integration? by steveha · · Score: 4, Interesting

    Whatever happened to wafer-scale integration?

    I read an article about this years ago. The idea was something like this:

    Memory chips are made on wafers. They are made side-by-side, then sliced apart, then each tested and mounted in a package. (Then eventually mounted on a little circuit board, and thence into our hands to install into our computers.)

    The idea was to make a wafer of memory chips, but not to just have them side-by-side; actually have traces connecting one chip to another. Then use the whole wafer as a RAM unit. You would need to test and find any defective RAM chips in the wafer, then cut a trace (or burn out a fuse, or whatever) to disconnect them from the rest of the wafer. (Not too different from bad-block management on a hard drive, really.) Finally you could make a stack of these wafers in a box, and sell it as a disk drive.

    This should be much cheaper than current RAM-based disk drives. It would be slower (the traces connecting the chips would be slower than a direct memory bus to each chip) but still way faster than a drive with moving parts.

    My understanding is that wafer-scale integration isn't very interesting for most applications, but for the specific niche of RAM-based storage units it seemed promising. Clearly I'm wrong since it didn't happen. Anyone know why?

    steveha

    --
    lf(1): it's like ls(1) but sorts filenames by extension, tersely
    1. Re:Wafer-scale integration? by sooth... · · Score: 1

      I have often wondered the same thing. I even worked for a facility capable of doing it too. They weren't interested.

      Frankly, I think they were narrow minded and that's why they are soon to go out of business.

      William...

    2. Re:Wafer-scale integration? by ddoctorisin · · Score: 1

      I agree. Although I think it may be that the wafers could suffer from some form of attenuation.

    3. Re:Wafer-scale integration? by andfarm · · Score: 1

      I think I heard something about this at one point -- the problem was apparently that chip failure rates are too high to make this work well.

      --

      TANSTAAFI: There Ain't No Such Thing As A Free iPod.

    4. Re:Wafer-scale integration? by sooth... · · Score: 1

      Right... Low voltage drops and the like as well.

      I was envisioning an EPROM type wafer, slower still I know but much faster still than HD access. This way though the data would survive a power down. CRCs could verify integrity on power on if needed.

      Even an inexpensive but small double sided 4.5" wafers would hold a substantial amount of data especially with 6 or more EPI layers on it.

      Wafers usually have bad areas on them, and that would be ok too. Just mark them off as bad and store the data allocation table in an off wafer EPROM for retrieval each time the medium was mounted by the OS.

      During the mounting process of the wafer to the power leads the entire wafer could be tested and mapped for good and bad sectors. Also, I imagine that there would be multiple mountings around the wafer to help power and transmission issues.

      Multiple wafers could be stacked in this manner to achieve more storage per bay. Greater data densities could be achieved when applying newer imaging processes.

      800mm wafers are a bit large for this type of application if it was ever intended to be placed into the consumer market though.

      The idea for using smaller wafers makes sense over large ones due to the through away nature of the product. One would be expecting some sectors to be blocked off anyway and imperfect wafers would be welcome so long as a certain percentage of them were viable. Flatness may be the only critical factor.

      The whole medium would have to be reasonably shock resistant as well, but that's not that hard to solve either by the mounting process.

      Next question, how to apply the circuitry to all the unique sectors? What about hard wired appliqué via a laserjet thermal transfer made specifically for each tested wafer?

      It's a pet project for someone more technical than me, but an interesting one non the less.

      William...
      be achieved by scaling to smaller circuts

    5. Re:Wafer-scale integration? by ddoctorisin · · Score: 1

      After much hunting... Well googling it I found something I feel may be applicable. Those of your electronics engineers or photonic engineers feel free to correct me. It appears there is a limitation on wafer sizes due to defects. As per this article "Creating large area FPGA's is limited by defective sections and the maximum reticule print size (~3x3 cm)". So I guess in theory if you were able to isolate the defects in a wafer or eliminate them and create a larger reticule print size. Then you could use the same principal for a solid state hdd made out of one giant wafer using redundant signal paths through the wafer. Or could you? G.H. Chapman, "FPGA Design for Decimeter Scale Integration (DMSI)", IEEE Intern. Symposium on Defect and Fault Tolerence in VLSI Systems, pg. 64-72, Austin 1998

    6. Re:Wafer-scale integration? by ddoctorisin · · Score: 1

      Or what if you kept within current limtations of wafer sizes whatever they may be 3x3cm or whatever. and just joined them together using some form of laser bridging

    7. Re:Wafer-scale integration? by flaming-opus · · Score: 1

      Where is the economic impetus for such a design? Any time you do custom work, it's not cheaper than commodity designs.

      An SSD can be made with an embedded microcontroller to handle the fibre channel interfaces, and aggregating the IO. The memory can be off-the shelf ECC RAM, you throw in some redundant power connections, multiple batteries, some off-the-shelf hard-drives and some clever firmware, and off you go. With devices like this you're not necessarily paying for the hardware, the cost comes from all the engineering time divided across a very small market.

      wafer-scale integration was once a holy grail for CPU designers, back when transister count was hard to get. By the time anyone got close to figuring out how to connect execution units in a way that you could tolerate having some not function, CPU design had moved on. It is no longer a lack of ALUs that hold back CPUs, but rather the coordination of pipelines. Now the challenge is keeping the pipelines full, due to high memory latency. Thus wafer scale integration is of marginal importance to CPU designers. (note bellow)

      As for memory applications, to be successful, wafer scale integration would have to offer something that a multi-chip solution can't offer. Since you have to limit your signal speed and electrical design to talk across the memory bus anyway, does it really hurt much to have a multi-chip modules (a DIMM) anyway? Yes it adds some to the physical packaging size, but it simplifies quality control, removes complexity, and is useable across the technology marketplace. Economies of scale are much more important than the component costs.

      One area where waffer-scale integration (sort of) works in the marketplace is cache. On modern CPUs, only a tiny fraction of the transistors are used for the actual processor, the majority of the chip is actually used for L2 cache. This is made of SRAM latches, so it is very bulky. On the athlon64 for sure, and probably on other chips, half the L2 cache can be disabled. This allows AMD to sell chips with a manufacturing defect as a "reduced-performance" chip. They drop the performance rating by 200 compared with the same speed chip with a full-sized cache. This is the basic idea of waffer-scale, just slightly more specific. As cache sizes increase, (they likely will, as chips are only becoming more vulnerable to RAM latency) this will likely become more important.

  51. Re:Performance and Cost by MrChuck · · Score: 4, Informative
    1. PCI @ 150MHz * 64bits is still 1200MB/s (that handwaves away any sort of overhead a bus might impose). So how are they getting 3036MB/s into a machine? Certainly not FCAL or SCSI. magic?
    2. When offered SSDs to work with for a project that needed lots of writes/reads of short lived files, I found them expensive, but good. I'm sorry to hear that Platypus ate it, I liked them and the idea of RAM on a PCI card - and they had linux and freebsd drivers. But SSDs COST a BUTTLOAD.
    3. I've used these guys' disks for years because they have tested to be as fast as SSDs. But with a half terrabyte behind them.

      With a battery backed cache of mirrored RAM, we found that for quick read/write stuff, the disks never got hit. If the data stayed, they ended up on the drives. If power was lost, the battery kept the cache alive for well over a day (I got bored and it met the "30 minutes" criteria we were looking at).

      The cache isn't huge (512? 256MB?) but it never filled. Basic elevator algorithms (we all did CS classes, right?) let the RAID side take data out of the cache in DISK order and write it out.

      And, not being Computer Vendor RAID, we found that it was fast and not expensive (given professional RAID). 15KRPM disks and dual controllers and dual PS and all that. Not for home use, but certainly for pro use. Oh and it gives great stats. Find stripe usage and cache hits on a Sun T3 that performs at half the speed for a good bit more money.

    I don't work for them, I just like their stuff. They're a small(ish) company that just does raid with lots of Wall St and corporate clients.
  52. Compact Flash (and the like) by jj00 · · Score: 1

    isn't the Compact Flash card I have sitting on my desk considered solid state? Why not just make a larger Flash card to replace a hard drive? Seems like the easier (and cheaper) way to go...

    1. Re:Compact Flash (and the like) by DarkSarin · · Score: 1

      Cheap? Do the math.

      Compact flash prices vary widely (.14 USD per MB for a 2GB module or .39 USD per MB for a 64MB model). Currently SATA HDD's cost as little as 0.0007625 USD per MB, by comparison (This is not an average, its just one drive -- a seagate 7200 RPM 160GB drive, so YMMV).

      Of course, compared to DDR RAM, compact flash may be cheaper, but for mass storage its not really an option.

      --
      "We don't know what we are doing, but we are doing it very carefully,..." Wherry, R.J. Personnel Psychology (1995)
    2. Re:Compact Flash (and the like) by LurkerXXX · · Score: 1

      Because the system described uses DRAM. DRAM is fast. Flash is SLOW. Very Very Very SLOOOOWWWW. DRAM access time is ~100 times FASTER than a hard drive. Flash is MUCH slower than a hard drive. Much.

    3. Re:Compact Flash (and the like) by Anonymous Coward · · Score: 0

      isn't the Compact Flash card I have sitting on my desk considered solid state?

      Yes.

      Seems like the easier

      Maybe.

      (and cheaper)

      No.

      Brought to you blah blah blah attention span.

    4. Re:Compact Flash (and the like) by warrax_666 · · Score: 1

      Just to add a bit to the other replies: Flash also has a very low limit on the number of total writes to any particular "cell" before that cell "dies". (IIRC it was something like 100.000, but at any rate, it was low).

      You can get around this by using alternative types of file systems which spread the writes evenly across the drive (I suppose a log-based file system would be close to ideal), but even so this is a serious limitation.

      --
      HAND.
  53. no crashes? by MrChuck · · Score: 1
    It's like skiing
    If you don't fall occasionally, you're not doing it hard enough.

    You clearly need to push those drives more.

  54. Reliability by odie_q · · Score: 1

    I've had one or two (I'm not sure about the second one yet, it could be something else in the system) drive failures at home. This is out of a great many hard drives I have owned. Several more have failed after laying about somewhere and not being used for a couple of years, but I don't count those.

    When I worked at a medium sized office (550 PC's) we would have roughly one drive failure a week, and a PC replacement cycle of 3 years (which would occasionally stretch to 4 years depending on the economic situation, but let's ignore that for now).

    So, If my tired mind is not playing tricks on me, that means 30% of the drives failed within their first three years of operation. A home computer is generally not used as much (it's the spinning up and down that wears most on the drive), and often has better quality components(!) than the utter crap HP/Compaq put in their machines these days.

    The interesting thing is that significantly more of these drives failed in their first year than in their third.

    Actually, the one failure a week number is probably slightly exagerated, the real value is likely closer to 20% failed within three years.

    --
    ...ceterum censeo Carthaginem esse delendam.
  55. The RAM itself is one thing... by Kjella · · Score: 4, Interesting

    ...that you can actually do in software (get a PowerMac or AMD-64, load it up with 8-12GB of ram). What usually cost money, is to have some sort of flush-to-disk feature.

    I'd love a SSD at least big enough to boot from, to combine with some other fanless stuff to create a 100% fanless, no moving parts PC (except from burner, which is silent when not in use). That + GbLan (to copy everything in from fanless machine, no damn spinning CD/DVD) using a direct crossover cable to a file server, preferably in a sound-isolated galaxy far, far away.

    That is my dream for my next setup. I've looked at doing the same simply dragging DVI + USB cables + external burner at machine, but it's not that great. A network cable can go so much longer...

    Kjella

    --
    Live today, because you never know what tomorrow brings
    1. Re:The RAM itself is one thing... by flonker · · Score: 1

      Have you looked into Boot from LAN? That way you can skip the SSD.

    2. Re:The RAM itself is one thing... by Anonymous Coward · · Score: 0

      I'd love a SSD at least big enough to boot from,

      You can get really cheap adapters to plug CF cards into an IDE bus. Get a 2GB CF card and run the OS with no swap.

    3. Re:The RAM itself is one thing... by DrSkwid · · Score: 1

      SSD to boot from ? It's called Compact Flash

      it comes from 1Mb to 512Mb and beyond.

      I boot FreeBSD and Plan9 without an HD

      and plan9 is ideally suited to diskless operation as it used a genuine dedicated file server that you can put anywhere on the network.

      So stick a nice noisy multi disk RAID system down somewhere you can't hear it and boot diskless from your 1Ghz EPIA workstation for a silent life.

      --
      There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter
  56. More Shitty Moderation! by Anonymous Coward · · Score: 0

    Congratulations mods! Keep up the censorship!

  57. Re:Performance and Cost by DigiShaman · · Score: 2, Informative

    Cenatek offers the Rocket Drive. Basically, it's RAM on a PCI card that requires an external power source via AC adapter. Of course, you could use a RAM drive setting in your OS. And if you need gigs of memory to start out with, your better off going with a 64bit platform as it scales.

    http://www.cenatek.com/product_rocketdrive.cfm

    --
    Life is not for the lazy.
  58. Graphics by Dark+Bard · · Score: 1

    I run into hard drive bottlenecks all the time working with large graphic files. They could be interesting for a "C" drive. 5 to 10 gig is adequate for operating system, basic software and the file you are working with. Just store your saved files on another drive. The big downside is safety. I haven't read the article yet but these systems rely on batteries to keep the ram active. If you loose power you loose the infomation and have to start from scratch. I've had hard drives setting on the shelf for over a year plugged them in and pulled off files. I'm sure the batteries have a long life but they are still another risk to data. If the ram drives get large enough, 100 gig+, they would be excellent for editing film. You could run full res 35mm captures real time. If the software was rewritten you wouldn't need to load a scene to edit it could be done right off the drive. As some one else pointed out the bus system on the motherboard would have to be improved as well. I guess the real question is there any reason the C drive can't be ultimately replaced by ram? Other than the fact most use Microsoft. Can you say reboot to clear ram?

  59. liked that movie by Anonymous Coward · · Score: 0

    Yeah, I liked that movie the first time I saw it, when it was called "Ram Disk."

  60. So, essentially... by Mr.+Neutron · · Score: 1
    ...it's a HDD with an absurdly huge cache:

    Uncompromised SDRAM data integrity is maintained through both battery backup and redundant disk drives, so your data is always protected.

    --
    dinner: it's what's for beer
    1. Re:So, essentially... by austus · · Score: 1

      It's not. Imagine a hard drive and a physical array of ram. Imagine further a circuit board containing an eprom/chip with firmware(burned-in software) that:

      1. turns that array of ram into a ram disk.
      2. manages mirroring of data between the ram disk and the hard drive.
      3. manages communication between the computer and the "hard drive" in a transparent way such that the computer has no clue it's dealing with a ram disk.

      Throw in a uninterruptable power supply of course to make sure the system always has the time to ensure the ram disk and harddrive are properly mirrored. Also throw in a fiber optic pin-compatible cable to replace the crap that passes for scsi/ide cables these days.

      This is an oversimplification of course, but I think is the gist of what is going on.

      The major advantages of this are twofold. The obvious benefit of such a system is speed. A less obvious benefit would be more reliability of the hard drive since it will experience vastly less usage since it would only be used during the mirroring process.

      In theory, this is a great solution because of the effective increase in speed and reliability. In practice, I don't know how this will ever be an inexpensive solution. How inexpensive can huge arrays of ram ever be?

  61. That would work fine by Mr.+Neutron · · Score: 3, Insightful
    until the janitor tripped over the power cable and all of your data goes *poof*

    It's not just the solid state memory, it's the ability to independently mirror it on disk as well. I suppose you could do that with software, but having that taken care of transparently and without CPU is a huge convenience factor.

    --
    dinner: it's what's for beer
    1. Re:That would work fine by Slashamatic · · Score: 1

      There is a CPU in there to manage the thing. This manages things like parttioning into logical drives and does the save/restore to HD.

  62. 64 GB RAM - Free Computer Included by Handpaper · · Score: 1
    Lots of connectivity, redundant power supplies. See here*

    * when fully configured

  63. Anybody thought about boot time? by Marxist+Hacker+42 · · Score: 0

    Say you do lose power with this thing. Sure- it's got backup on the platter drives. Triple Redundant too- enabling odd-man-out error correction. So for the first SEVERAL minutes, you're restoring the RAM from the hard disks, doing three reads at a time, comparing the result, and placing it back into RAM.

    This alone will keep us from *ever* seeing a consumer version

    But dangnabit, if I was running a huge data center I'd want a RAED (redundant array of expensive disks) of these puppies- two alone mirrored would create 100% uptime easily. Lose power to one, replace it's power supply, and run off the other until the first one boots again.

    --
    SJW: a person who perceives an injustice, and while correcting it, commits a greater injustice.
  64. 3GB/sec disk transfer - 3kbs web transfer by NynexNinja · · Score: 1

    These guys need a faster web site.

  65. Re:Performance and Cost by Anonymous Coward · · Score: 0

    What happened to Platypus? Those guys had an 8Gb Solid-state disk that plugged into a PCI slot.....

  66. Rebadged TMS RAM SAN by lazn · · Score: 2, Informative
    am I the only one to notice this is just a rebadged Texas Memory Systems RAM-SAN?

    http://www.superssd.com/default.asp

    ==>Lazn

  67. Re:Performance and Cost by JamieF · · Score: 1

    OK, so they put some memory sockets on a PCI card, made an ASIC that pretends to be an IDE controller or something, and they charge $999.00 for it.

    No wonder I haven't heard of them before.

  68. DRAM!!! by the_2nd_coming · · Score: 1

    ARE THEY NUTS!!! power outage...woops there goes my paper, and my OS.

    how about ust giving a hard drive a 32 MB cache? or more perhaps?

    that would be cheaper and be more reliable until MRAM hits the market and can replace Hard drives with no worries, not to mention no more hybernation by writing to disk, you can just turn off the computer and it will boot right back up to where you left off in a few seconds.

    --



    I am the Alpha and the Omega-3
    1. Re:DRAM!!! by Your+Average+Joe · · Score: 1

      hey, clue: raid cards turn off disk cache for reliable disk transactions...

      --
      Your Average Joe
  69. POINTLESS (use system RAM instead) by Theovon · · Score: 1

    How can this possibly be better than just putting more RAM in to your computer. Ok, sure, if you have a 32-bit machine, then you're stuck. But if you have the money for this beast, than you have a 64-bit server box, and you can install 64MB of RAM.

    This would be EVEN FASTER because you don't have to get your disk data from over some SCSI bus or whatever. Instead, you just rely in your OS to do a good job of caching disk blocks in RAM. If all of your data fits in RAM, then the only disk access that will ever happen is writes, and that'll be so well buffered that it'll never impact anything. (Just be sure to use a journalling file system!)

  70. huh? by djcatnip · · Score: 3, Funny

    It uses DRAM memory to store data instead of spinning platter hard drives, giving an access time of just 20 microseconds.' It still does use platter-based drives but it's a cool idea anyway.

    <jon stewart> Whaaaaaa? </jon stewart> How can you use DRAM instead of platters but still be platter based?

    Oh, yes. Now I understand. Way to be crystal clear there. I know the formula, (submit well written story with exacting details that clearly illustrate the point minus "well written" and minus "with exacting deatils that clearly illustrate the point"), and yet... I keep taking the bait.

    and let me beat you to the punchline:

    "you must be new here..."

    "you realize you're on slashdot..."

    (Score: -1 Duh)

    --
    I make these: http://beatseqr.com
    1. Re:huh? by Slashamatic · · Score: 1

      The DDRAM is battery backed, but there is a small HD there which backs up the solidstate memory so when power fails, it will use the battery for a set period and then copy the contents to HD to preserve it.

  71. Re:Performance and Cost by Anonymous Coward · · Score: 0
    Basic elevator algorithms (we all did CS classes, right?)
    for those who don't know what kind of algorithm is that :
    http://www.cs.wisc.edu/~solomon/cs537/disksched .ht ml
  72. Re:Performance and Cost by Anonymous Coward · · Score: 0

    The same link web browser friendly :

    the link :P

  73. To anyone who has bought / implemented an SSD... by JamieF · · Score: 2, Insightful

    Why is this so much better than a bunch of additional system memory?

    Theoretically if you throw a bunch of RAM into a computer with a remotely modern OS, some of that memory will be used to buffer writes, and performance will improve. An exception would be in cases where applications go out of their way to force writes all the way to disk, such as in databases with their transaction logs.

    Is the problem just that there are so many applications that sync() all the time, that a hardware buffering solution such as SSD is required because otherwise the OS's file buffers are constantly being flushed? (Yes I understand that SSD is persistent, or at least much more persistent than plain old RAM that dies when the power goes off.)

    I can see the enterprise-friendly angle of "just add this disk and the whole system goes much faster" instead of trying to rewrite existing apps or tune the hell out of the OS. I'm just curious about particular cases where adding RAM doesn't work but SSD works well... are there enough to justify the existence of these devices?

  74. Re:Performance and Cost by Anonymous Coward · · Score: 0

    Ok, one that works better this time :
    grrr

  75. Re:To anyone who has bought / implemented an SSD.. by Anonymous Coward · · Score: 0

    Because as a FC based device, which looks like spinning storage
    such devices can be used as the back end storage for large
    clustered file systems, i.e. data/files can be accessed from
    a variety of systems, MPP, PVP, etc.

  76. Re:To anyone who has bought / implemented an SSD.. by ColdZero · · Score: 0

    Because not all processes will continue while something is still in RAM. They will wait until they are written to disk before going on. In those cases, they will still be waiting for that poor little platter to spin around. The oodles of ram in the computer won't make a difference.

  77. ugh its just VFS by sPaKr · · Score: 1

    Ugh, moving VFS out of system memory and into dedicated memory on the HD doest make it any better. Its still not perminate storage. And when you do hit that memory critical you swap out memory to memory.. instead of just skipping the swap. It seems like this is really targed against PAE or 64bit chips. Bah I like my memory where it belongs swaping out early just becouse the OS doest know that its just a one big memcpy is LAME.

    1. Re:ugh its just VFS by Anonymous Coward · · Score: 0

      Wow, congrats on proving how incompetent you are. What do you do if your OS crashes and you are running a mission critical database on it? If the data is on a SAN unit you can do clustered file systems and failover and just keep running, if it's in your machine, you just lost everything, no time to flush the last transactions to disk.

  78. This is hardly new by Nom+du+Keyboard · · Score: 1
    It uses DRAM memory to store data instead of spinning platter

    You know, 25+ years ago they had solid state paging drums for the mainframes of the time. Sure they were larger in size, less capacity, slower than this unit, and more expensive -- but so were computers in general. This is hardly a new idea.

    --
    "It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
  79. Re:FP by Trejkaz · · Score: 1

    Hot damn! Gimme gimme gimme!!

    --
    Karma: It's all a bunch of tree-huggin' hippy crap!
  80. A solution via Cat5 by Grakun · · Score: 1

    You can use a KVM Extender to set the keyboard, mouse, and monitor several rooms away from the computer. Many of them also do serial, audio, and usb. Just get one that uses Cat5 cable between the transmitter and receiver.

  81. MRAM by anethema · · Score: 1

    I cant wait for MRAM solid state disks. Non-volatile fast, dense...etc

    Has all the makings of a great future hard drive replacement.

    Small sizes of it are already available, at least on paper.

    Datasheets here.

    --


    It's easier to fight for one's principles than to live up to them.
  82. Re:Performance and Cost by DigiShaman · · Score: 1

    Ya the price is lame, I know :(

    But if your after a true SSD in both IDE and SCSI configurations, check out the stuff M-Systems has to offer. They even come packaged in the same form factor as a standard HD.

    http://www.m-sys.com

    --
    Life is not for the lazy.
  83. Why do they call it solid state? by vegetasaiyajin · · Score: 1

    Aren't conventional hard drives solid too?

    --

    My heart is pure, but make no mistake, it's pure evil
  84. Re:Performance and Cost by Anonymous Coward · · Score: 0
    if you need gigs of memory to start out with
    Well it's not swap or VM (and yeah, I've been using 64bit systems for 9 years when I got an Alpha in). It's fast storage that survives reboots.

    For applications that need quick read/write access like a, er, MAIL SPOOL, this is perfect. If my main box just spools up to 4 Gig of RAM, then fine. If it's full, there are secondaries (and fallbackMX hosts in sendmail can push slower mail off to another machine).

    Wonder if drivers for various un*xes are available. Or even (heaven forbid) documentation needed to write drivers for those dead OSs like BSD.

  85. Re:Performance and Cost by Anonymous Coward · · Score: 0

    about that PREVIEW button that we all have...

  86. Re:Performance and Cost by Anonymous Coward · · Score: 0

    I know a guy who works for BlueArc. They do most of the network and filesystem stuff in hardware, hardly anything is firmware. (the rare filesystem permission checks or something)

    20Gbps is only 16% short of the 3GBps metioned in this article, and you get up 256 TB. Really expensive shit, but it's pretty cool.

  87. This Story is about 14 years late by anonymous+cupboard · · Score: 1
    This must be a new record for /.!!! The original Eurex is based in Frankfurt, Germany and has been running since 1989 when it was known as DTB. Eurex is a financial futures and options exchange,

    The backend runs OpenVMS and the critical files are kept on solid-state disks. Originally DEC ESE20s. In particular, the order-books are kept on mirrored solid-state disks to allow for fast matching between buyer and seller.

  88. Trading by Slashamatic · · Score: 2, Insightful
    If you read the article you find that a financial exchange uses this. It is essentially running a double continuous auction in about 30,000 products (you must runs separate books for calls and put options and then for each exercise price and series). Futures require one per delivery. It doesn't take a lot of basic products to generate thousands of actual things that can be bought and sold.

    For each of these items there is a price/time linked list of bids and offers which are used to determine a market. This could be done in memory but if you pull the plug on the system, then the order-book dissappears. This is why they put it onto disk. However with a requirement of subsecond response to any of the several thousand participants - high performace disks are a must.

  89. I have one of these SSDs........ by Anonymous Coward · · Score: 0

    I own a 4GB SSD Rocketdrive. First i set it up to host my "temporary internet files",my %TEMP% environment variable, and my pagefile. I've got a Athlon 64 3.2ghz system with 1 gig of PC3200 ram. I didn't notice the system running any faster. Even when i used the drive as my photoshop scratch disk it seems about as fast as my raid 0 WD raptors (36gb's). I'd say save your dough unless your machine is a real POS and then yeah maybe one of these would help.

  90. Since when are ramdrives news? by p3d0 · · Score: 1

    Is there actually something new here?

    --
    Patrick Doyle
    I mod down every jackass who puts his moderation policy in his sig. Oh, wait a sec....
  91. Thank you... by Kjella · · Score: 1

    ...for pointing out a non-solution to a non-problem. Unless you have a need for *one* master version (e.g. database), it is pointless. It's like using a Cray to render a movie.

    Kjella

    --
    Live today, because you never know what tomorrow brings
  92. Re:Performance and Cost by MrChuck · · Score: 1
    Ya the price is lame, I know :(

    I have a $950 2.5GB barracuda I bought. I hooked it up to the home SPARC 2 and moved 4 535MB disks onto it. And turned off those disks. External with a little soft gasket work between the disk and the frame, it was pretty quiet. AT least from the next room over (bed).

    $1k for bus speed disk is GREAT.

    The M-Systems (disk on chip) and CF disks and those are nice for, say, storing configs, but not live data. A CF is typically faster than a floppy, but can be slower that a CD (and is much slower than a disk drive).

    Now, can we get a file system that journals onto a big bit of NVRAM disk? If I have a 1GB NVRAM card let reiserfs cache meta data and even files to it. And in the background (after the fsync()), move the data from that to the drive as time allows).