Slashdot Mirror


A Hybrid Approach For SSD Speed From Your 2TB HDD

Claave writes "bit-tech.net reports that SilverStone has announced a device that daisy-chains an SSD with a hard disk, with the aim of providing SSD speeds plus loads of storage space. The SilverStone HDDBoost is a hard disk caddy with an integrated storage controller, and is an easy upgrade for your PC. The device copies the 'front-end' of your hard disk to the SSD, and tells your OS to prefer the SSD when possible. SSD speeds for a 2TB storage device? Yep, sounds good to me!"

194 comments

  1. Save your money... by damn_registrars · · Score: 1, Informative
    To use this in a desktop, you need
    • An available 3.5" bay
    • A 2.5" hard drive
    • An SSD of whatever size you can afford

    This seems like a lot of money to spend for potentially not a lot of speed. Generally, 2.5" hard drives aren't quite as fast as their 3.5" counterparts anyways, so you're spending a fair bit of money to speed up something that wasn't really made for speed anyways.

    Sure, you can "drop it right in" to your existing computer, assuming that your desktop is for some reason already using 2.5" SATA drives. And if your desktop is currently using 2.5" SATA drives you probably didn't build it to be a speed demon anyways.

    --
    Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
    1. Re:Save your money... by GungaDan · · Score: 2, Insightful

      I don't see where a 2.5" HD is required - 3.5" should be fine. The gizmo looks like a 2.5" to 3.5" adapter tray, but the HD is not installed in the gizmo.

      Besides, have you ever heard of a 2.5" 2TB drive?

      --
      Eloi are stupid, throw morlocks at them!
    2. Re:Save your money... by TheRaven64 · · Score: 5, Informative

      Or, you can just use ZFS and turn on the L2ARC, which will use the SSD as a cache for the hard disks and not need any custom hardware.

      --
      I am TheRaven on Soylent News
    3. Re:Save your money... by kungfuj35u5 · · Score: 1

      Or, you can just use ZFS and turn on the L2ARC, which will use the SSD as a cache for the hard disks and not need any custom hardware.

      Exactly what I was thinking, I think you may have even beaten me to the post. This seems like something a good operating system can implement with relative ease without even needing a custom tailored filesystem to do so. Not that that doesn't help.

    4. Re:Save your money... by adonoman · · Score: 2, Interesting

      Or just plug in a usb drive into any Windows 7 computer.

    5. Re:Save your money... by Anonymous Coward · · Score: 1, Informative

      I don't see where a 2.5" HD is required

      If you RTFA, you'll see

      The device takes the form of a 2.5in to 3.5in hard disk caddy with a couple of SATA connectors on the end

      Which makes sense, as a 3.5 bracket with a 3.5 hard drive and an SSD would not fit in a standard 3.5 bay.

      It may be possible instead to dedicate one of your 3.5 bays to this, running SATA cables from your 3.5 drive to it (and then to the SATA controller). But then you've just used a full bay for one SSD and a silly gizmo.

      Besides, have you ever heard of a 2.5" 2TB drive?

      No, but we don't let reality get in the way of a good slashvertisement around here.

      If only they made 2.5" SSD's then you could put THAT int he caddy and use a regular 3.5" in another bay. If only such magical things existed.

    6. Re:Save your money... by lorax · · Score: 1

      According to the pictures, you put a 2.5 inch SSD in the caddy and daisy chain it to an additional 3.5 inch HD in a different bay

    7. Re:Save your money... by Anonymous Coward · · Score: 4, Informative

      USB drives speeds are in the 20-30 MB/s. SSD drives are 150-250 MB/s. Conventional HDDs are 50-100 MB/s

    8. Re:Save your money... by jgagnon · · Score: 3, Informative

      Except that then you're at USB speeds instead of SATA speeds.

      --
      Remember to maintain your supply of /facepalm oil to prevent chafing.
    9. Re:Save your money... by zippthorne · · Score: 0, Troll

      Great Idea. ... Computer management, clickly clicky, drop list.. NTFS or FAT. Which one of those is ZFS?

      Let's try another option.
      $ mkfs.zfs /dev/disk/by-label/MyMainDisk

      No command 'mkfs.zfs' found, did you mean:
        Command 'mkfs.gfs' from package 'gfs-tools' (main)
        Command 'mkfs.hfs' from package 'hfsprogs' (universe)
        Command 'mkfs.bfs' from package 'util-linux' (main)
        Command 'mkfs.xfs' from package 'xfsprogs' (main)
        Command 'mkfs.ufs' from package 'ufsutils' (universe)
        Command 'mkfs.jfs' from package 'jfsutils' (main)
      mkfs.zfs: command not found

      Okay...hmm.

      $ diskutil listFilesystems | grep -i zfs

      hm, nothing.

      I suppose it *should* be possible to do in software, and I'd even imagine that like RAID, the benefits of doing it in hardware become more dubious as tome goes on. However, there is the question of the least effort way of getting it done in "my" computer right now.

      --
      Can you be Even More Awesome?!
    10. Re:Save your money... by iamhassi · · Score: 2, Interesting

      True, and everything sounded find until I read this
      "Every time the system starts, HDDBOOST will initiate mirror backup automatically to ensure front-end data between the two drives are the same."

      on every system start it's going to create a mirror backup, which sounded bad unless it works like Mirror RAID which doesn't take any time at all, it mirrors in real-time

      So basically on every startup it mirrors the HD to the SSD, then pulls everything from the SSD until it needs to write data. Writing data goes on the HD, not the SSD. When starting up again the SSD mirrors all the new written data from the HD and continues on.

      Sounds technically feasible and that should be faster and I'd love to see some benchmarks although I'm not sure how that'd work because reading data should look incredibly fast since it's on the SSD but written data goes to the HD so that'd be normal speed meaning a benchmark might not show amazingly fast speeds yet your PC should be noticeably quicker.

      --
      my karma will be here long after I'm gone
    11. Re:Save your money... by ByOhTek · · Score: 2, Insightful

      If you RTFA you'd find the 2.5" drive is for the SSD, not the rotational drive.

      The bracket mounts the SSD inside of it, and then passes failed requests to the HDD, which is external to the bracket.

      --
      Self proclaimed typo king, and inventor of the bear destroying coffee table (patent not pending).
    12. Re:Save your money... by 0100010001010011 · · Score: 2, Informative

      Who the hell creates a ZFS drive by using mkfs.zfs?

      zpool create tank mirror sdb sdc works just fine under Debian.

    13. Re:Save your money... by Anonymous Coward · · Score: 1, Insightful

      You place too much value on a single drive bay. Buy a larger case already.

    14. Re:Save your money... by yabos · · Score: 1

      Um, do you think this device is aimed at anyone that would know how to set up a ZFS pool like that? All you have to do with this thing is plug in 2 drives and install a driver if their claims are true. It's only 33 Euros which is not a bad price, although who knows if they'll come out with drivers for other OS'

    15. Re:Save your money... by Pojut · · Score: 1

      Screw what is required to do it, I'm more concerned about trying to find a 2TB drive that has decent reliability. So far, every 2 TB hard drive I have looked at has had a lot of problems, at least according to online reviews from multiple sources (forums, newegg, IT guys I know, etc.) Obviously, the people who complain tend to be the loudest, but still...I haven't come across a single 2TB drive that didn't have a LARGE number of complainers.

      The same can't be said for 1TB or 1.5TB drives.

    16. Re:Save your money... by AmberBlackCat · · Score: 2, Funny

      I'm bookmarking your post so I have something, to link to, the next time somebody asks why I prefer the GUI over the command line.

    17. Re:Save your money... by Anonymous Coward · · Score: 0

      You do realize that of the three operating systems I mentioned, the one you have a solution for is the most tinkerable and least commonly used desktop OS.

      Also, the package that provides zpool is zfs-fuse. The "use" stands for user. As in, it runs in user space rather than kernel space. Which is good, because it's apparently not even on version 1.0 yet. And it's only in sid. And it can't have been even there for very long because it's not in Karmic.

      So.. yeah.. zfs.. not really ready or available for general use yet.

    18. Re:Save your money... by 0100010001010011 · · Score: 1

      zfs-fuse> is not available for general use.

      I have no problem trusting my data to ZFS in FreeBSD or OpenSolaris. (Where the commands to create a zfs 'file system' are the same.

    19. Re:Save your money... by illumin8 · · Score: 1

      Or, you can just use ZFS and turn on the L2ARC, which will use the SSD as a cache for the hard disks and not need any custom hardware.

      Is anything like L2ARC available for Linux? I would love to have something like this in our database servers.

      --
      "When the president does it, that means it's not illegal." - Richard M. Nixon
  2. Just a cache? by Erich · · Score: 4, Insightful

    Haven't disk manufacturers been doing this forever, using faster memories to cache disk? I guess the difference now is that the memory is slower than DRAM and non-volatile so it isn't lost in the event of power failure? Or maybe you can get more flash storage at a low price point?

    --

    -- Erich

    Slashdot reader since 1997

    1. Re:Just a cache? by girlintraining · · Score: 1, Funny

      Haven't disk manufacturers been doing this forever, using faster memories to cache disk? I guess the difference now is that the memory is slower than DRAM and non-volatile so it isn't lost in the event of power failure? Or maybe you can get more flash storage at a low price point?

      "DRAM is yesterday's news. SSDs are the future. And they cost a lot more than a pile of memory chips, therefore they must be better. Order fifty." -- Some PHB

      --
      #fuckbeta #iamslashdot #dicemustdie
    2. Re:Just a cache? by eln · · Score: 1

      High-end storage devices have been using SSD for years to speed things up. It basically allows for a larger cache than RAM (for less money), and also means non-volatile cache like you noted. Of course, how much of a speed gain you get depends on what your workload looks like and how good their caching algorithms are. So, I'm not impressed at all by this little device, but I would be impressed if it came with a new and more efficient caching algorithm.

    3. Re:Just a cache? by Drethon · · Score: 1

      I have a Lenovo laptop that has something like 2G flash drive doing basically that. So I'd say its probably pretty much just a cache...

    4. Re:Just a cache? by asdf7890 · · Score: 2, Insightful

      "DRAM is yesterday's news. SSDs are the future. And they cost a lot more than a pile of memory chips, therefore they must be better. Order fifty." -- Some PHB

      SSDs are cheaper per Gb then DRAM. Compare the price of a 64Gb SSD to 64Gb with of decent DDR2/3 RAM. The per-Gb price difference is likely to grow too as SSD prices are falling more sharply than DRAM prices at the moment.

    5. Re:Just a cache? by Sandbags · · Score: 0

      Of course, you don't need DDRII/DDRIII RAM, as even the worst performance, cheapest, high capacity sticks you can get are still much faster than SSDs.

      But actually, he said DRAM, not SDRAM or DDR SDRAM, but actually I think he means NAND, since SSD drives in fact ARE DRAM!

      --
      There is no contest in life for which the unprepared have the advantage.
    6. Re:Just a cache? by Anonymous Coward · · Score: 0

      Yes, it's been called a cache, and also a buffer.
      It's even been done with non-volatile memory before as well.
      This is just the latest version to come out, seems more 'universal' than many of the others, and has a much better price point than previous ones.
      (I remember one from the late 80s made with volatile ram that cost around a $1000. Worked beautifully, but way to expensive for the public.)

    7. Re:Just a cache? by amorsen · · Score: 2, Informative

      since SSD drives in fact ARE DRAM!

      No. DRAM doesn't keep state across power loss, and it's a lot more expensive, just like the OP said.

      --
      Finally! A year of moderation! Ready for 2019?
    8. Re:Just a cache? by Anonymous Coward · · Score: 0

      Once upon a time, SSDs were DRAM-based, but for some time now they have been based on Flash, usually NAND Flash.

      SSDs aren't DRAM.

    9. Re:Just a cache? by should_be_linear · · Score: 1

      DRAM can only be used as read-cache. SSD is much larger and is read and write cache.

      --
      839*929
    10. Re:Just a cache? by Sycraft-fu · · Score: 1

      The main problem with HDD cache is it is so small with regards to the space it is caching. Yes, HDs have cache but 16MB and 32MB are common. There are a couple 64MB ones but that's it. That's for 500GB-2TB. Can't cache much data with that. Compare that to your CPU which tend to have somewhere in the 4-8MB of L2/L3 cache and usually RAM in the 2-8GB range. There's a lot less oversubscription of the cache which makes it works better. A system like this has the potential to work a lot better since you can equalize the difference much better. You can have 64GB for a 2TB drive, instead of 64MB.

      Also yes, non volatility matters. You can't go and have a massive RAM cache on a disk if it means something could be written to that, power could fail, and then it never gets committed. Something like this you can write to it as often as you like and that can be delayed for disk write as long as you like because it is non volatile.

      That is another area that kills disk performance is doing simultaneous reads and writes. Disk needs to read a bunch of data, but also needs to do some small writes during that. Well, would make sense to delay the writes and batch them, but you don't want to do that or you lose data. No problem here, flash is fast for random access so reads and writes at the same time should be a big deal.

    11. Re:Just a cache? by BikeHelmet · · Score: 1

      DRAM cache mostly helps with reading - not writing. I guess that's where the SSD comes in. Our filesystems just aren't set up to allow writing huge swathes of tiny files sequentially, so when random write speeds are important the SSD is king.

      Although I suppose ZFS might be set up for exactly that... but few of the others are.

      An HDD with a huge cache would be interesting. One company tried it a few years back.

      http://arstechnica.com/hardware/news/2007/09/coming-soon-hard-drives-with-1gb-ddr-ram-cache.ars

      Based on this 5400RPM drive's performance, I imagine a WD Black(7200RPM with dual heads) would come close to maxing out SATA2 if it had a gigabyte or two of cache.

    12. Re:Just a cache? by Eil · · Score: 1

      Yeah, it's just a cache. But it's a really big cache.

      Actually, it sounds a lot to me like a beige-box version of Intel's Turbo Memory thing for laptops, which only has drivers for Windows.

  3. Sounds like bullshit to me by Anonymous Coward · · Score: 1, Interesting

    How would the disk supposedly know, which part of the 2TB I am going to need next?

    1. Re:Sounds like bullshit to me by obarthelemy · · Score: 2, Interesting

      My guess is, it won't. It must be a dumb cache that just monitors which sectors/clusters are most often read, and caches those.

      It may be better than the current use of SSDs though, which is to put a whole OS on them even though there's many parts of the OS that are barely used. If I had an SSD, IE8 would be on it !

      --
      The Cloud - because you don't care if your apps and data are up in the air.
    2. Re:Sounds like bullshit to me by izomiac · · Score: 1

      Essentially this is a high capacity "bulk storage" drive and a fast "working" drive. So the potential customers seem to be either those who are too lazy to copy their data from "storage" to "working" manually, or those so predictable that that a computer program can cache stuff before they ask for it. My guess is that the predictive algorithm is basically "the user asked for this a second ago, let's copy it to the SSD in case the user has a short term memory deficiency".

      OTOH, this would help with some stupid programs that are unable to cope with multiple drives/partitions. E.g. my last laptop's TV tuner insisted upon storing everything on an NTFS D:\, so either I could map D:\ to a USB harddrive and not use it when I was traveling, or limit myself to the few gigabytes I had to spare on my laptop's internal drive.

  4. 2.5" drives only by gEvil+(beta) · · Score: 0, Troll

    The device takes the form of a 2.5in to 3.5in hard disk caddy with a couple of SATA connectors on the end.

    Good job Claave! You apparently didn't even get to the second paragraph before submitting the article. You can't use a 2TB hard drive with this because there are no 2TB 2.5" drives yet.

    --
    This guy's the limit!
    1. Re:2.5" drives only by gEvil+(beta) · · Score: 1

      Derrrr. Nevermind. It helps to also look at the pictures when commenting on TFA.

      --
      This guy's the limit!
    2. Re:2.5" drives only by Cocoronixx · · Score: 1

      Also helps to comprehend WTF you are reading. The excerpt you quoted from the article states the availability of a 3.5in hard disk caddy for chrissakes!!?

      --
      "Obscenity is the crutch of the inarticulate motherfucker." - cloak42
    3. Re:2.5" drives only by NitroWolf · · Score: 0, Offtopic

      The device takes the form of a 2.5in to 3.5in hard disk caddy with a couple of SATA connectors on the end.

      Good job Claave! You apparently didn't even get to the second paragraph before submitting the article. You can't use a 2TB hard drive with this because there are no 2TB 2.5" drives yet.

      Good job, GEvil! You didn't even read the article at ALL!

      Please point to where it says you must use a 2.5" hard drive. Hmm, you don't think that maybe... JUST MAYBE... the 2.5" caddy is FOR THE SSD so that you can mount it in a 3.5" bay? Then you mount your 3.5" HD as normal!

      Gosh... reading comprehension! Learn it. Live it! Love it! Take it home and call it George.

      kthxbye

    4. Re:2.5" drives only by gEvil+(beta) · · Score: 1

      Also helps to comprehend WTF you are reading. The excerpt you quoted from the article states the availability of a 3.5in hard disk caddy for chrissakes!!?

      Read it again and look at the pictures. The caddy is the size of a 3.5" drive, in which you install a 2.5" SSD. That is what they mean by a "2.5in to 3.5in hard disk caddy".

      --
      This guy's the limit!
    5. Re:2.5" drives only by gEvil+(beta) · · Score: 1, Offtopic

      Hence the "Derrr nevermind" post I made six minutes before you posted this diatribe.

      --
      This guy's the limit!
    6. Re:2.5" drives only by Cocoronixx · · Score: 5, Funny

      Read it again and look at the pictures. The caddy is the size of a 3.5" drive, in which you install a 2.5" SSD. That is what they mean by a "2.5in to 3.5in hard disk caddy".

      AGAIN!? How dare you accuse me of reading TFA. Are you trying to ruin my /. cred?

      --
      "Obscenity is the crutch of the inarticulate motherfucker." - cloak42
    7. Re:2.5" drives only by NitroWolf · · Score: 1, Offtopic

      Hence the "Derrr nevermind" post I made six minutes before you posted this diatribe.

      Once again, you are full of fail. You don't think that maybe that was written just AFTER you made your uninformed post and PRIOR to when you decided to actually read the article and then post a correction?

      Gosh, I dunno... maybe. We aren't in an IRC chat room. I know that's kind of hard to grasp for people with short attention spans, but really... we aren't. This is a message forum, it's not real time.

    8. Re:2.5" drives only by Em+Emalb · · Score: 3, Funny

      LOL, nerd fight. Someone call a doctor, they might get a paper-cut.

      --
      Sent from your iPad.
    9. Re:2.5" drives only by DavidRawling · · Score: 1

      Thankfully, in the continuing search for safety, paper will be outlawed in favour of bits and bytes in our new super-safe nerd haven. Oh, won't somebody think of the children^W nerds?

  5. Pick the false statement by sakdoctor · · Score: 4, Insightful

    No software or driver update is required

    Some software is needed to achieve the magic

    1. Re:Pick the false statement by Seth+Kriticos · · Score: 4, Informative

      They are a bit confusing. The manual ( http://www.silverstonetek.com/downloads/Manual/storage/Multi-HDDBOOST-Manual.pdf ) though says, that the HDD has to be de-fragmented before usage. They don't mention other software, though they mention Windows here and there. The manual states though, that any OS supporting SATA will do.

      (It's a Windows pussy thing again, you can freely ignore it).

    2. Re:Pick the false statement by mariushm · · Score: 2, Informative

      From what I read in the article, you have to defragment the big hard drive because the SSD will fill up with the data at the beginning of the drive. The read and write requests are just caught by whatever chip they use so when the OS requests a read from the start of the drive, the data from SSD is sent instead. Same for writes.

      So the SSD acts just like level 2 cache for the first 32-64-whatever GB of the big disk. The rest will never be optimized.

    3. Re:Pick the false statement by complete+loony · · Score: 3, Informative

      Yeah, it simply mirrors the start of your HD, up to the size of the SSD. It doesn't look like there's any intelligence in the way it caches content. So if you have more data stored on the HD than the size of the SSD, the performance of reading this data will be ever so slightly slower than reading from the HD directly.

      --
      09F91102 no, 455FE104 nope, F190A1E8 uh-uh, 7A5F8A09 that's not it, C87294CE no. Ah! 452F6E403CDF10714E41DFAA257D313F.
    4. Re:Pick the false statement by Fred_A · · Score: 3, Funny

      So if you have more data stored on the HD than the size of the SSD, the performance of reading this data will be ever so slightly slower than reading from the HD directly.

      Obviously you need a 2TB SSD in order to get the maximum performance gains from the device.

      --

      May contain traces of nut.
      Made from the freshest electrons.
    5. Re:Pick the false statement by jimthehorsegod · · Score: 1

      But potentially still useful for a clean install - since the system and apps will therefore be amongst the first xGB...

    6. Re:Pick the false statement by komode0 · · Score: 1

      You RTFA AND RTFM? My internet hat is off to you, sir.

  6. You mean like in... by kungfuj35u5 · · Score: 5, Informative

    ZFS? Hybrid storage pools have been around for a long while, and exist as a pretty well balanced software solution to this problem. Hybrid solid-state/magnetic disks were in the market as well which used a similar technique. There is nothing new or impressive about this device.

    1. Re:You mean like in... by Demonantis · · Score: 1

      I usually just put the OS on a scsi and data on a separate drive(possible raid if desired). This is not really interesting news unless the guy does it at a hardware level.

  7. What 2TB HD? by damn_registrars · · Score: 0, Redundant

    This adapter is for 2.5" hard drives - if you put a 3.5 drive in it, you wouldn't fit drive+adapter+SSD into a 3.5" bay. Who makes a 2TB 2.5" SATA drive currently? I am not aware of any...

    --
    Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
    1. Re:What 2TB HD? by NitroWolf · · Score: 4, Informative

      This adapter is for 2.5" hard drives - if you put a 3.5 drive in it, you wouldn't fit drive+adapter+SSD into a 3.5" bay. Who makes a 2TB 2.5" SATA drive currently? I am not aware of any...

      Seriously... did not one read the article? You mount the fucking 2.5" drive in the caddy and mount your 3.5" HD where you would normally mount it and run a fucking cable from your HD to the caddy. Is this so fucking hard to get a grasp on? For christs sake.

  8. Holy carp! by NitroWolf · · Score: 1

    Well... it looks like there finally might be a reason to spend the money on an SSD. Up until now, it would be a nice speed boost, but the cost:performance ratio is so out of whack for SSDs, it just makes purchasing one ridiculous unless you have some very specific needs. For 95% of the people who have purchased them, they just want the biggest e-peen. That's fine and all, but my days of swinging around the biggest e-peen are over, so I've held off buying an SSD until the prices drop and capacity goes WAYYYY up.

    However, with this particular device, it actually makes it worth it to spring for a lower capacity, fast SSD (for naturally less money than the higher capacity ones) that will cache the files I use the most. The question is, and it wasn't really clear from the article unfortunately, is it a real time "mirror" - in so far as over time, if I start using more file and others less, will the drive start caching those newer files that I use more than the older ones I am using less? Assuming it does (since it would be kind of useless if not), this makes an 80 GB SSD a viable option!

    However, the one drawback I see to this is my current RAID 0 setup would be unusable and I'd have to switch back to using one drive. That's not a terrible thing, as I've never been too thrilled with the whole RAID 0 thing and if the minor speed advantages it imparts are fully mitigated by the SSD - switching over to a single 2TB drive is awesome.

    I would definitely shell out some bucks for this solution, assuming it works as advertised.

    1. Re:Holy carp! by radish · · Score: 2, Insightful

      Well... it looks like there finally might be a reason to spend the money on an SSD. Up until now, it would be a nice speed boost, but the cost:performance ratio is so out of whack for SSDs, it just makes purchasing one ridiculous unless you have some very specific needs

      "Very specific needs" like wanting my OS & apps to load as fast as possible? Putting OS, apps, pagefile etc on the SSD greatly improves system responsiveness. FLACs, MP4s & JPGs can stay on a spinning disk, I don't need to access them so quickly. A couple hundred bucks on a smallish SSD gives you a MUCH better performance kick that spending the equivalent on RAM or CPU, in my experience (provided of course you have at least an average spec machine to start with).

      --

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

    2. Re:Holy carp! by Phleg · · Score: 1

      RAID 0 is for chumps. You get a similar read speed boost from RAID 1, and you don't have the dramatically increased risk of failure.

      --
      No comment.
    3. Re:Holy carp! by Anonymous Coward · · Score: 0

      Only a fucking dipshit would use RAID 0 for a system drive.

    4. Re:Holy carp! by obarthelemy · · Score: 1

      Holy carp indeed... always loved that fish symbolism for the Christians.

      --
      The Cloud - because you don't care if your apps and data are up in the air.
    5. Re:Holy carp! by timeOday · · Score: 2, Informative

      RAID 0 is for chumps. You get a similar read speed boost from RAID 1*

      * read performance only

    6. Re:Holy carp! by evilviper · · Score: 1

      Very specific needs" like wanting my OS & apps to load as fast as possible? Putting OS, apps, pagefile etc on the SSD greatly improves system responsiveness. FLACs, MP4s & JPGs can stay on a spinning disk,

      Putting your swap file on a RAM-Disk has long been the stereotypical geek example of human stupidity...someone who knows just enough to be very dangerous.

      Additionally, if you have excess amounts of RAM available, every modern operating system will cache all disk reads, thereby offering instant access to your apps/files, the SECOND time you open them. Combine this with eg. S3/suspend support proliferating, greatly reducing the reasons for most people to ever shutdown their systems, and with the fact that system memory caching will be faster than any disk, and it's not hard to make the case that more RAM is more beneficial for most common usage patterns.

      --
      Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
    7. Re:Holy carp! by ftobin · · Score: 2, Insightful

      While that is true in many cases, when one drive dies you will get much better read and write performance out of RAID 1 than RAID 0.

    8. Re:Holy carp! by petermgreen · · Score: 1

      Putting your swap file on a RAM-Disk has long been the stereotypical geek example of human stupidity...someone who knows just enough to be very dangerous.
      Putting a swapfile on a regular ramdisk (that uses up system memory) is indeed stupid.

      Putting swap on storage that is faster than a HDD but not directly usable as ram seems pretty sensible to me in certain situations (admittedly I don't belive those certain situations are the average users desktop).

      One thing i've learnt is that the cost of a system has a highly nonlinear relationship to the ammount of ram required with small increments in cost until you get to the limit of your chosen family of computers then a large jump as you move to a higher end family.

      Additionally, if you have excess amounts of RAM available, every modern operating system will cache all disk reads, thereby offering instant access to your apps/files, the SECOND time you open them.
      And then some virus scan runs overnight and kicks everything out of ram so you have to wait for it to grind back in when the morning comes.

      The problem is our desktops are running on operating systems and software designed with a batch processing mindset. Sure there has been the odd tweak for interactivity but we still rely on the OS to guess what it should cache in memory and what it shouldn't rather than designing knowlage of deadlines and reuse probabilities into our software.

      --
      note: i'm known as plugwash most places but i screwd up registering that here somehow in the past and now can't register
    9. Re:Holy carp! by NitroWolf · · Score: 1

      "Very specific needs" like wanting my OS & apps to load as fast as possible? Putting OS, apps, pagefile etc on the SSD greatly improves system responsiveness. FLACs, MP4s & JPGs can stay on a spinning disk, I don't need to access them so quickly. A couple hundred bucks on a smallish SSD gives you a MUCH better performance kick that spending the equivalent on RAM or CPU, in my experience (provided of course you have at least an average spec machine to start with).

      Haha... I don't know if you're trying to be serious or funny. If serious, then I kind of feel sorry for you, but your e-peen is huge, man, huge!

      You quoted me saying "Very specific needs" and within 2 words, you use the word "want."

      I said absolutely nothing about wants in the quoted text. Sure, I WANT a huge SSD array. However, I do not need one. Sure, you WANT your OS and apps to load as fast as possible, but you do not need it. Very few people NEED the speed of an SSD. The cost:performance ratio is outside the realm of reality for most people, you included I'm sorry to be the bearer of that bad news, but you do not NEED your OS and apps to load that extra 2 or 3 seconds faster than a mechanical hard drive. Your life and livelihood are not going to be affected if your OS loads in 23 seconds vs 38 seconds. If your MS Word document takes 3 seconds to open up instead of 1, your world will not come crashing down and your customers are not going to leave in droves. You do not NEED the speed of an SSD.

      A couple hundred bucks will net you almost 4 TB of space vs the 128 - 256 GB of space the same money of an SSD drive will net you. You may not need 4 TB of space and 128 GB may be more than adequate for your needs... but we get back to the cost:performance ratio, which is what my whole post was about and which you failed to apparently understand. The cost:performance ratio, as measured against a mechanical hard drive is not there except in some very rare circumstances that very few people in the world will encounter at the present time; You are one of these people. I am also one of these people.

      Now, if you want to talk about swinging your SSD e-peen around, as I said, you win my good man, you win. If this product performs as advertised, I will spend under $100 for the same performance you get and have 10x the space... at that point, you lose man, your e-peen just fell off!

    10. Re:Holy carp! by Bryan+Ischo · · Score: 1

      Give it up man. Your points are inherently biased by your own cost:performance:size relationship values. Saying that SSDs have no value to anyone because they don't fit your own usage models is just RETARDED. I'm not even going to explain to you my usage models, or the other usage models like mine, for which an SSD has proven to be a very nice, and well worthwhile, upgrade, worth every penny. I'm not going to explain it because it's not necessary for me to use examples to make the very simple point that I am making: not everybody wants the same thing you do, so stop with the ridiculous generalizations and RETARDED references to "e-peen", which is just about the most juvenile phrase I have read in quite a long time.

    11. Re:Holy carp! by radish · · Score: 2, Informative

      Putting your swap file on a RAM-Disk has long been the stereotypical geek example of human stupidity...someone who knows just enough to be very dangerous.

      Likewise declaring someone stupid when it turns out YOU are the one who needs to do a little learning. Quoting the Windows Engineering Blog:

      Should the pagefile be placed on SSDs?

      Yes. Most pagefile operations are small random reads or larger sequential writes, both of which are types of operations that SSDs handle well.

      In looking at telemetry data from thousands of traces and focusing on pagefile reads and writes, we find that

              * Pagefile.sys reads outnumber pagefile.sys writes by about 40 to 1,
              * Pagefile.sys read sizes are typically quite small, with 67% less than or equal to 4 KB, and 88% less than 16 KB.
              * Pagefile.sys writes are relatively large, with 62% greater than or equal to 128 KB and 45% being exactly 1 MB in size.

      In fact, given typical pagefile reference patterns and the favorable performance characteristics SSDs have on those patterns, there are few files better than the pagefile to place on an SSD.

      Read that last sentence to yourself a few times, let it sink in. Now you can say sorry.

      Additionally, if you have excess amounts of RAM available, every modern operating system will cache all disk reads, thereby offering instant access to your apps/files, the SECOND time you open them

      Indeed useful. I get the last 2-3GB's of accessed files at RAM speed, if I'm lucky. That doesn't help boot time, that doesn't help sleep time, and that doesn't help when I launch an app for the first time in a while. All of which ARE helped greatly by using an SSD. Why can't I have both?

      and it's not hard to make the case that more RAM is more beneficial for most common usage patterns.

      Yeah, except I have one, and you're wrong. Caching is great for files you hit a lot, but you know what? My system drive has 40-50GB on it, and adding 40-50GB of RAM isn't really an option for most people, certainly not an economical one. You'll be lucky to get 10% of that in the cache - VERY lucky (seeing as all those media and data files will be pushing out the useful stuff you actually stand a chance of wanting to read again).

      You seem to have a problem with SSDs, that's great, don't buy one. I wouldn't trade mine for 16GBs of RAM, never mind 8 (which is the equivalent cost).

      --

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

    12. Re:Holy carp! by radish · · Score: 1

      I said absolutely nothing about wants in the quoted text. Sure, I WANT a huge SSD array. However, I do not need one

      I think it's very unlikely you NEED a computer at all. Let's say you do, does it NEED to be faster than a 486sx25? Or do you just WANT to be able to compile something in less than an hour? When people talk about computer "needs" it's usually understood that we mean "needs to fulfill some desire", because very few people can claim to need any kind of computation device in the same way they need food, water or shelter.

      I don't NEED 4TB of disk space (well, not an extra 4TB anyway), as you correctly point out. And no I don't NEED my system to boot in half the time it used to, or firefox to come up in 1 second vs 15. The point I was trying to make is that the $200 I spend on an SSD made more difference to the day-to-day general performance of my machine than the $500 I spent upgrading the CPU and memory. So it's better value as an upgrade, IMHO. That runs contrary to most people's common sense (as demonstrated by this thread) making it all the more important to try and spread the word as far as I'm concerned.

      --

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

    13. Re:Holy carp! by NitroWolf · · Score: 1

      Give it up man. Your points are inherently biased by your own cost:performance:size relationship values. Saying that SSDs have no value to anyone because they don't fit your own usage models is just RETARDED. I'm not even going to explain to you my usage models, or the other usage models like mine, for which an SSD has proven to be a very nice, and well worthwhile, upgrade, worth every penny. I'm not going to explain it because it's not necessary for me to use examples to make the very simple point that I am making: not everybody wants the same thing you do, so stop with the ridiculous generalizations and RETARDED references to "e-peen", which is just about the most juvenile phrase I have read in quite a long time.

      You really don't have much in the way of reading comprehension, do you? My statements are not biased in any way, shape or form. They are merely statements of fact. Please quote where I said that SSDs have no value to anyone, or heck, even quote where I said they have no value. You can't, because I said just the opposite. They have a value and that value is different to different people. For the vast majority of people, it is a status symbol for the perceived differences in their day to day work. As I said, loading your Word document in 1 seconds as opposed to 3 seconds is not a huge deal. If that's worth $600 to you, great - I did not say that was wrong. I merely stated that the cost:performance ratio outside a reasonable realm for most people.

      Since, by your post, you are asserting that the cost:performance ratio is NOT outside the realm of reasonable for most people, I would like you to point to some statistics that show SSDs have replaced mechanical hard drives, either in existing installations or in new sales. In fact, why are they still selling mechanical hard drives? Oh... it may be because the cost:performance ratio of a mechnanical hard drive is more reasonable.

      There's no generalizations here. I said the majority of people do not NEED an SSD. Of those that say they NEED an SSD, the majority of those do not actually NEED one but WANT one.

      So again, your post is completely devoid of any useful context as your whole argument is based off of things I did not say or imply. Way to go in setting up that straw man and good job in knocking it down!

    14. Re:Holy carp! by NitroWolf · · Score: 1

      I said absolutely nothing about wants in the quoted text. Sure, I WANT a huge SSD array. However, I do not need one

      I think it's very unlikely you NEED a computer at all. Let's say you do, does it NEED to be faster than a 486sx25? Or do you just WANT to be able to compile something in less than an hour? When people talk about computer "needs" it's usually understood that we mean "needs to fulfill some desire", because very few people can claim to need any kind of computation device in the same way they need food, water or shelter.

      I don't NEED 4TB of disk space (well, not an extra 4TB anyway), as you correctly point out. And no I don't NEED my system to boot in half the time it used to, or firefox to come up in 1 second vs 15. The point I was trying to make is that the $200 I spend on an SSD made more difference to the day-to-day general performance of my machine than the $500 I spent upgrading the CPU and memory. So it's better value as an upgrade, IMHO. That runs contrary to most people's common sense (as demonstrated by this thread) making it all the more important to try and spread the word as far as I'm concerned.

      If you want to get pedantic about the word NEED, then you are correct. However, I qualified my definition of NEED in this context by saying:

      Your life and livelihood are not going to be affected if your OS loads in 23 seconds vs 38 seconds. If your MS Word document takes 3 seconds to open up instead of 1, your world will not come crashing down and your customers are not going to leave in droves.

      I'm not talking about base human needs, I'm talking about the needs of business or day to day life. The fact is, you do not NEED your Word documents to open in 1 second vs 3. You do not NEED your OS to boot in 23 seconds as opposed to 38.

      There are people who do, an example would be a car computer that needs to boot up as fast as possible (I'm not talking about an ECU here, I'm talking about something like an aftermarket entertainment system or some such) - if the computer is slow to boot, people would not buy that product, whereas if it was on and running by the time they put the car in drive, they'd be happy with it. That is a NEED. People will not buy the product because it's too slow vs. buying the product because it is fast.

      Another example is a high capacity data gathering system ... it may NEED to write the data at 180 MB/s for a sustained period.

      You and the majority of the people who feel they NEED an SSD drive are not in this category. They are in the WANT category.

      There is NOTHING WRONG WITH THIS. I am not stating that there is, nor have I stated that there is in previous posts, which you and others seem to be unable to grasp. My original post said that purchasing one is ridiculous unless you have very specific needs, then you go on to list your wants. I submit, again, that it's ridiculous given the alternatives available. It may be acceptable to you to spend the money for the performance but that makes it no less ridiculous. People spend money on ridiculous things all the time, and again there is NOTHING WRONG WITH THAT. I do it myself. But to say that spending $600 to save 2 seconds on occasion and to save another 20 seconds on the occasions that you actually reboot or power on is reasonable is insulting to the majority of the world population.

      I am glad you have that sort of disposable income and I am glad that _I_ have that sort of disposable income. However, I do not kid myself into believing that buying status symbols, such as SSDs, is anything but buying a status symbol when a mechanical hard drive will provide 75 - 90% of the performance at 1/8th the price.

    15. Re:Holy carp! by Bryan+Ischo · · Score: 1

      If your entire point revolves around a semantic quibble about the meanings of the word 'need' versus 'want', then I've got nothing left to say, as this sort of argument is worthless.

    16. Re:Holy carp! by evilviper · · Score: 1

      In fact, given typical pagefile reference patterns and the favorable performance characteristics SSDs have on those patterns, there are few files better than the pagefile to place on an SSD.

      Read that last sentence to yourself a few times, let it sink in. Now you can say sorry.

      I fully agree that a pagefile on a solid-state disk will be much faster than a hard disk. You must also agree that a page-file on a RAMDISK would be even faster... Both express a fundamental misunderstanding, however.

      A) I'm not sorry.
      B) I did not call you an idiot... but your response isn't exactly impressive.

      That doesn't help boot time,

      Indeed not, which is why I pointed out that people are finding it necessary to reboot less and less these days.

      that doesn't help sleep time

      I fail to see your point. Sleep (S3 Suspect) is almost instantaneous, and will see no speed improvement from solid-state storage.

      You'll be lucky to get 10% of that in the cache - VERY lucky (seeing as all those media and data files will be pushing out the useful stuff you actually stand a chance of wanting to read again).

      There are methods for both Unix systems and more recent Windows systems to force the most frequently accessed files to remain cached in RAM, NOT allowing "media and data" to push it out.

      You seem to have a problem with SSDs

      No, merely your justification of them, and the false logic on display here.

      --
      Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
  9. Isn't this just a fancy cache? by Anonymous Coward · · Score: 1, Insightful

    I DNRTFA, but this really just seems like a fancy version of cache to me....

    1. Re:Isn't this just a fancy cache? by Skuld-Chan · · Score: 1
    2. Re:Isn't this just a fancy cache? by argent · · Score: 2, Insightful

      No, it's a simple version of cache that doesn't actually do proper caching. All it does is preloading, and only over part of the device. Most of the volume of the hard disk will have no performance boost at all. You'd almost certainly be better off just having two devices, and using junction points on Windows or soft links on UNIX to move the frequently accessed files to the smaller disk.

    3. Re:Isn't this just a fancy cache? by complete+loony · · Score: 1

      While I agree that you might be better off managing what is on your SSD yourself just by treating it as a separate disk drive. With a good 3rd party defragmenter, this could actually be useful.

      For example My Defrag can be scripted to move files into specific places on your drive based on just about any criteria. So you could set it up to move your OS, Program, Temp, Swap, Hibernate and any other recently accessed files to the start of the disk. Then move everything else to the end.

      --
      09F91102 no, 455FE104 nope, F190A1E8 uh-uh, 7A5F8A09 that's not it, C87294CE no. Ah! 452F6E403CDF10714E41DFAA257D313F.
    4. Re:Isn't this just a fancy cache? by argent · · Score: 1

      Or you could save your money and make the SSD your system drive and the rotating media your archive.

      I'd wait for version 2 or 3, the one that actually uses the SSD as a block level cache with your choice of replacement policies.

  10. look at the picture accompanying TFA - it uses 3.5 by GungaDan · · Score: 1

    This thing uses 3.5" drives, or you could slap it onto a 2.5" drive if you wanted to. The thing TAKES THE FORM OF a drive caddy - it is not a drive caddy.

    --
    Eloi are stupid, throw morlocks at them!
  11. Re:look at the picture accompanying TFA - it uses by gEvil+(beta) · · Score: 2, Informative

    This thing uses 3.5" drives, or you could slap it onto a 2.5" drive if you wanted to. The thing TAKES THE FORM OF a drive caddy - it is not a drive caddy.

    Actually, after looking at it more, it is a drive caddy -- for a 2.5" SSD. This device basically acts as a daisy chain controller that you hook both a 2.5" SSD and a regular 3.5" HD to. The controller then presents the combined device to the BIOS/OS as a single drive.

    --
    This guy's the limit!
  12. Windows Only by asdf7890 · · Score: 2

    In order to appear as one storage device in Windows, SilverStone has needed to use some software to...

    There is the turn off for me. If I were to use something like this I would want an OS agnostic solution. Of course that would mean the caching would have to be done at the block level rather than the file level so it might not be able to be as bright (a block level cache manager wouldn't know to deallocate space on the SSD immediately when a file is deleted for instance), but it should be quite practical to design an algorithm that keeps the most often used blocks in the cache (the SSD) without the whole thing being needless wiped first time you copy a massive data file in (you wouldn't want that 20Gb file to be written to the SSD first time it is laid down, at the expense of dropping blocks frmo OS startup files and such, in case it is hardly ever accessed again - for instance an image of a blueray disc that you are copying to another disc would not want to touch the cache as it'll probably be written one, read once then wiped. How this block-based cache management algorithm would work in detail is left as an exercise for the reader...

    1. Re:Windows Only by meustrus · · Score: 1

      Well, just based on the line you quoted, I would say that the software is so that the two hard drives only show up as one hard drive. It says nothing about file systems, or caching, or file/block level, it just says that they made this neat piece of software that will eliminate your confusion as to which of the two drive letters you should use. I'm guessing that they didn't make such a neat little piece of software for other OSes, in which case, if I'm reading it right, it should still work but show two disks.

      --
      I sometimes ask revealing, often ignorant-seeming questions. Maybe they're harder to answer than you think.
    2. Re:Windows Only by asdf7890 · · Score: 1

      if I'm reading it right, it should still work but show two disks.

      At which point you migth as well just plug the two drives directly into your main I/O controller and not bother with the extra device...

    3. Re:Windows Only by Anonymous Coward · · Score: 0

      There is the turn off for me. If I were to use something like this I would want an OS agnostic solution. Of course that would mean the caching would have to be done at the block level rather than the file level so it might not be able to be as bright (a block level cache manager wouldn't know to deallocate space on the SSD immediately when a file is deleted for instance), but it should be quite practical to design an algorithm that keeps the most often used blocks in the cache (the SSD) without the whole thing being needless wiped first time you copy a massive data file in (you wouldn't want that 20Gb file to be written to the SSD first time it is laid down, at the expense of dropping blocks frmo OS startup files and such, in case it is hardly ever accessed again - for instance an image of a blueray disc that you are copying to another disc would not want to touch the cache as it'll probably be written one, read once then wiped. How this block-based cache management algorithm would work in detail is left as an exercise for the reader...

      I can't imagine not wanting to use Windows so badly that I end up with no solution. And worse, when someone else doesn't want to use Windows and tells me "it is my honor" to figure it out. If God wanted me to write anything more complicated than a Bubble Sort then FSM wouldn't have given us NewEgg.

    4. Re:Windows Only by Anonymous Coward · · Score: 0

      a block level cache manager wouldn't know to deallocate space on the SSD immediately when a file is deleted for instance

      That is what the TRIM command is for on drives and OSes that support it. However only 7 fully supports it now and only when the hard drive reports a RPM of 0 (indicating an SSD). There are apps that can compute and send TRIM commands in other OSes, but these are not automatic upon deletion obviously. Upon file deletion, a TRIM command is sent to the drive indicating where the file used to be, telling the drive that nothing useful is stored there and it can be "garbage collected". This mainly is useful in that the built-in wear leveling or other algorithms does not have to "save" that block of data when it is doing housekeeping - it can be overrwritten without saving the contents.

      Certain filesystems would benefit from this more than others. For instance, the journal in journaling filesystems, and the "Hot Zone" of HFS+ would benefit from the additional access speed boost.. The problem is that these areas are updated very frequently, which is not good for flash. In other words, the uses for which flash would be best utilized are hardest on the flash.

      I also vote for huge DRAM caches instead.

  13. Re:look at the picture accompanying TFA - it uses by hattig · · Score: 1

    This stuff all needs to be within the same hard drive.

    Sell a 2TB 3.5" hard drive with a 64GB *fast* SSD cache on board (transparent to the OS), and then we will be talking.

  14. Waste of money and data safety by KDN · · Score: 2, Insightful
    Would it not be more cost effective to add more main memory to the machine? Main memory would be a lot faster than SSD ram. Also I have a concern that frequently updated blocks (like your file system superblocks) would not get written out to disk in a timely fashion.

    Now, maybe you could do it safely if the device had RRD ram to handle the caching, SSD flash ram to handle power outages, a rechargable battery or ultra cap to provide power to write the RRD ram to flash ram after a power outage, and a controller to handle all this. You would need to implement all the normal os buffer caching and writebacks as well.

    1. Re:Waste of money and data safety by drinkypoo · · Score: 2, Insightful

      Would it not be more cost effective to add more main memory to the machine?

      I've never had a machine with capacity for more than 12GB RAM.

      Also I have a concern that frequently updated blocks (like your file system superblocks) would not get written out to disk in a timely fashion.

      That's a bigger problem with your RAM-based approach than with a flash-based one.

      Now, maybe you could do it safely if the device had RRD ram to handle the caching, SSD flash ram to handle power outages, a rechargable battery or ultra cap to provide power to write the RRD ram to flash ram after a power outage, and a controller to handle all this.

      You can get this with ZFS, a DRAM disk with a backup battery (volatile disks do exist that you load with DIMMs) and a caching raid controller with a backup battery.

      --
      "You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
    2. Re:Waste of money and data safety by Seth024 · · Score: 1

      This would provide a speed boost to machines that already have enough memory (and most do). Part of your operating system would be copied to the SSD (persistent storage!). A system boot would be much faster because the data can be accessed from the SSD and doesn't have to be read from the HDD.
      Having more memory, which only solves capacity problems, wouldn't be helpful in this case.

  15. dear kind sir by damn_registrars · · Score: 1, Funny
    Dear kind sir

    I suspect we may have read different articles. Your assertion of

    You mount the fucking 2.5" drive in the caddy and mount your 3.5" HD where you would normally mount it

    Does not match the article when it says

    The device takes the form of a 2.5in to 3.5in hard disk caddy

    So I thank you for your kind concern of

    un a fucking cable from your HD to the caddy. Is this so fucking hard to get a grasp on? For christs sake

    And I hope you have a very nice day kind sir.

    --
    Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
    1. Re:dear kind sir by Sandbags · · Score: 1

      The 2.5" caddy is for the SSD and includes the controller card. A traditional drive bay is required in ADDITION to this caddy (which fits a second 3.5" bay adjacent to the HDD).

      The man with the foul mouth is correct.

      total internal PC space required is 2 x 3.5" bays. You can absolutely use a 2TB drive and a small SSD with this kit.

      --
      There is no contest in life for which the unprepared have the advantage.
    2. Re:dear kind sir by NitroWolf · · Score: 0, Offtopic

      Dear kind sir

      I suspect we may have read different articles. Your assertion of

      You mount the fucking 2.5" drive in the caddy and mount your 3.5" HD where you would normally mount it

      Does not match the article when it says

      The device takes the form of a 2.5in to 3.5in hard disk caddy

      So I thank you for your kind concern of

      un a fucking cable from your HD to the caddy. Is this so fucking hard to get a grasp on? For christs sake

      And I hope you have a very nice day kind sir.

      Dear kind sir,

      You are a fucking idiot and can not read. Please do not continue to use the Internet, as even a short article is apparently beyond your comprehension.

      You even go so far as to quote the point that refutes your assertion, trying to use it as an example of why you are correct. This is, perhaps, one of the ultimate forms of demonstration of your rank stupidity. Please destroy your computer now, or at the very least disconnect your computer from the internet so that you do not harm other people with your base ignorance.

    3. Re:dear kind sir by Anonymous Coward · · Score: 0

      Please explain, from your point of view, what is going on in the top-right corner of this picture from the article.

    4. Re:dear kind sir by obarthelemy · · Score: 1

      It's a caddy that goes into a 3.5" bay, and accepts a 2.5" SSD. Your regular spinning HD goes into its normal 3.5 bay, and connects to the caddy.

      Thank you for your attention. Can we now go on and discuss the interesting things about that gizmo ?

      --
      The Cloud - because you don't care if your apps and data are up in the air.
    5. Re:dear kind sir by Anonymous Coward · · Score: 0

      You are being a bit of a cockbag about this all over the thread. Not saying you're wrong - the way you're expressing your opinion just has a bit of a cockbag feel to it. People generally don't take well to being abused so it's unlikely your attitude is going to inspire them to change their ways.

    6. Re:dear kind sir by Anonymous Coward · · Score: 0

      Dear kind sir

      I suspect we may have read different articles. Your assertion of

      You mount the fucking 2.5" drive in the caddy and mount your 3.5" HD where you would normally mount it

      Does not match the article when it says

      The device takes the form of a 2.5in to 3.5in hard disk caddy

      So I thank you for your kind concern of

      The 2.5in SSD goes into the devece; the 3.5in HDD goes somewhere else.

    7. Re:dear kind sir by damn_registrars · · Score: 2, Funny
      Oh joy of joys, I was hoping you would be so kind as to bestow unto me the gift of another kind response.

      You are a fucking idiot and can not read.

      Clearly, you are the truest source of all wisdom. You are a gentleman of caliber previously unattained. I am not worthy of your kindness, dear sir and I bow to your civility.

      Please do not continue to use the Internet, as even a short article is apparently beyond your comprehension.

      You have graced us with so much kindness in so little space.

      one of the ultimate forms of demonstration of your rank stupidity

      Why thank you kindly sir. I hope you could be so kind as to continue to heap praise upon me.

      Please destroy your computer now, or at the very least disconnect your computer from the internet so that you do not harm other people with your base ignorance

      I am so glad that you devoted so much time to discussing the topic at hand. A lesser man might have resorted to slinging insults rather than actually talking about the (lack of) technical merit of the device that this slashvertisement is dedicated to. I am infinitely thankful to you for being so kind and considerate.

      The depths of the technical aspects of this conversation are seldom reached or surpassed these days, I thank you for your immense insight and kindness in this matter.

      --
      Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
    8. Re:dear kind sir by NitroWolf · · Score: 0, Troll

      Oh joy of joys, I was hoping you would be so kind as to bestow unto me the gift of another kind response.

      You are a fucking idiot and can not read.

      Clearly, you are the truest source of all wisdom. You are a gentleman of caliber previously unattained. I am not worthy of your kindness, dear sir and I bow to your civility.

      Please do not continue to use the Internet, as even a short article is apparently beyond your comprehension.

      You have graced us with so much kindness in so little space.

      one of the ultimate forms of demonstration of your rank stupidity

      Why thank you kindly sir. I hope you could be so kind as to continue to heap praise upon me.

      Please destroy your computer now, or at the very least disconnect your computer from the internet so that you do not harm other people with your base ignorance

      I am so glad that you devoted so much time to discussing the topic at hand. A lesser man might have resorted to slinging insults rather than actually talking about the (lack of) technical merit of the device that this slashvertisement is dedicated to. I am infinitely thankful to you for being so kind and considerate.

      The depths of the technical aspects of this conversation are seldom reached or surpassed these days, I thank you for your immense insight and kindness in this matter.

      tl;dr. np.

    9. Re:dear kind sir by damn_registrars · · Score: 1, Funny

      tl;dr

      Indeed you are a scholar and a gentleman. I will cherish this wonderful conversation for the rest of my days and then some. I hope you a lifetime of peace and happiness for the kindness you have bestowed upon me and indeed the whole of the world. We all could never be adequately grateful for a kind soul such as yours.

      --
      Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
  16. Move along, nothing to see here folks. by Velorium · · Score: 1

    Come back when there are some benchmarks to look at.

  17. File system by olau · · Score: 1, Insightful

    There was a paper some years ago about building the file system in such a manner that smaller files were placed on an SSD ( 1 MB) and large files were placed on a harddisk. At that time, SSDs were a lot smaller than today though.

    Generally, it can make sense to discriminate your files because they don't all have the same space and access characteristics. Maybe 100 files is taking up 90% of the space compared to the other 9900 files. Maybe it's similar for the access pattern.

    Still, for the idea to fly, you need to a robust algorithm and it needs to be clever about the strengths of the hardware. For instance, SSDs aren't so hot at random writes, sadly. Less than 0.1 msec write time would be neat for an ACID database.

    1. Re:File system by Simetrical · · Score: 1

      For instance, SSDs aren't so hot at random writes, sadly. Less than 0.1 msec write time would be neat for an ACID database.

      Cheap SSDs aren't. There are now SSDs by Intel and OCZ that can do tens of megabytes per second of random writes. They just cost several times as much per gigabyte as any other disk out there.

      --
      MediaWiki developer, Total War Center sysadmin
  18. For those that didn't RTFA by HannethCom · · Score: 3, Informative

    This solution uses two 3.5 inch drive bays in your computer, one for your large platter drive, the other for the caddy with a SSD drive.

    Some software is installed (Windows only) that makes the two drives look like one.

    The most used files from the large drive are copies to the smaller SSD drive. When files cached on the SSD drive are requested, they are read from there, if they do not exist there the request is passed onto the bigger drive. If the file is being used enough it will be copied to the SSD drive at the same time as the information is getting sent to the computer. You will not get SSD drive speeds in this case.

    Yes, this is just using a SSD drive as a cache.

    The product does not come with SSD storage, you have to buy a SSD drive of your choosing as well as this caddy.

    --
    Microsoft, Apple, Google, Amazon what's the difference? All steal money from devs and control with walled gardens.
    1. Re:For those that didn't RTFA by Mashiki · · Score: 1

      Really it sounds and I RTFA, like an extension of RAID0.

      --
      Om, nomnomnom...
    2. Re:For those that didn't RTFA by mariushm · · Score: 1

      That's not how I see it. You have to defragment the big drive first, to have all executables and OS files at the start of the big disk. Then you use the software they give you to fill up the SSD with the data at the start of the big disk and after a reboot, reads and writes to those first 32 GB or so of the big disk will pass through the SSD.

    3. Re:For those that didn't RTFA by Zebra_X · · Score: 1

      "Microsoft has a rocket launcher pointed at their feet and they think they can rocket jump."

      Well, it worked in Marathon...

  19. Shouldn't this be integrated into the controller? by blackketter · · Score: 1

    It seems to me that the natural evolution in hard drives would be to build the flash cache on to the controller board on the hard drive. Is any drive manufacturer building this kind of hybrid flash/magnetic drive?

  20. RTMFA already by Weaselmancer · · Score: 4, Interesting

    They even have a fucking picture.

    The 2.5 caddy is for your SSD. Mount your 3.5 wherever you like.

    --
    Weaselmancer
    rediculous.
    1. Re:RTMFA already by damn_registrars · · Score: 0, Flamebait

      The 2.5 caddy is for your SSD. Mount your 3.5 wherever you like.

      So the caddy is just a collossal waste of space and money, then. You should be able to fit a 2.5 HD with a 2.5 SSD into a single 3.5 bay. Instead you're going to use a 3.5 bay for your SSD, so that you can achieve some minor performance gains on your regular HD?

      Why not just buy a caching HD controller instead?

      --
      Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
    2. Re:RTMFA already by Weaselmancer · · Score: 1

      I'm not going to defend the product. I don't care if you hate it. But please - hate it for the right reasons at least.

      You have now apparently read the article (or at least looked at the picture), and you now hate it for the right reasons.

      My job here... is done.

      --
      Weaselmancer
      rediculous.
    3. Re:RTMFA already by damn_registrars · · Score: 0, Flamebait

      You have now apparently read the article (or at least looked at the picture)

      I read the article - and looked at the picture - before my first post on this topic.

      and you now hate it for the right reasons.

      I don't really hate it. I dislike it. It has not wronged me so significantly that I would use the word hate as a description of how I feel about it.

      I previously disliked it for being a waste of space and money. I now dislike it for being a waste of space and money.

      --
      Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
    4. Re:RTMFA already by Anonymous Coward · · Score: 0

      I am curious - are you trolling, or do you not understand?

      The device requires TWO 3.5" bays.

      The first bay takes the gizmo, but you put a 2.5" SSD (most SSDs today are 2.5") into it first. That's the "caddy" part of it. There is no room in the gizmo for a second drive.

      The second bay takes a regular 3.5" drive, such as the oft-mentioned 2Tb drive. (You could put a 2.5" drive there, but you'd need a "caddy" for it)

      Yes, you could use a caching disk controller, but:

      1. the cache on the controller will be substantially smaller than an SSD - less cache
      2. Caching disk controllers take up a PCI or PCIe slot - these are usually fewer in number than 3.5" disk bays in the average PC
      3. This device plugs in to a regular SATA controller, and can respond to your normal disk controller software - caching disk controllers usually require special drivers

    5. Re:RTMFA already by damn_registrars · · Score: 0, Flamebait

      The device requires TWO 3.5" bays.

      That is exactly the problem I have with it. The device itself is a waste of physical space inside a computer case.

      Well, it wastes space and money, to be more specific...

      caching disk controllers usually require special drivers

      I don't know what kind of ultra-shitty walmart-grade caching controller you have in mind. A quality caching controller (like the ones DPT made back in the day) is completely OS-agnostic. No special drivers - or drivers of any sort - required. A good hardware caching controller does its job in hardware.

      --
      Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
    6. Re:RTMFA already by adolf · · Score: 1

      Did you expect the extra componentry to somehow consume zero space?

    7. Re:RTMFA already by damn_registrars · · Score: 0, Flamebait

      Did you expect the extra componentry to somehow consume zero space?

      I think there are more intelligent ways to consume physical space than this. For example, if we look at Hard Disk Form Factors we see that in the physical space allocated for a single 3.5 hard disk drive bay (4 x 1.63 x 8 inches) we can easily fit two 2.5 hard disk drives (2.75 x 0.374-0.59 x 3.945 inches each). Instead they waste the space by placing only a single 2.5" (or smaller) SSD into a 3.5" bay. And being as most 3.5" drives are now only 1" tall (in a 1.63" tall bay), there is enough space to place the SSD on top of the HDD and still fit the space.

      And there is history of adapters taking up (essentially) zero space - look for 80pin SCA to 68pin SCSI adapters. And if you want to cache your HDD access, a hardware caching controller could well be a more intelligent choice (and would be OS-agnostic as well).

      --
      Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
  21. Good operating system compatibility? by tepples · · Score: 1

    This seems like something a good operating system can implement with relative ease without even needing a custom tailored filesystem to do so.

    "A good operating system" isn't compatible with thousands of proprietary third-party apps and device drivers designed for the majority operating system for home PCs. Nor can the data on a drive set up for use with "a good operating system" easily be transplanted into another PC whose operating system is not aware of the look-aside SSD cache.

    1. Re:Good operating system compatibility? by kungfuj35u5 · · Score: 2, Interesting

      There are a number of options: Firstly: Use a different machine with samba if you want to give the storage to your windows hosts. This option isn't for everybody, but over gigabit while it will be a third of the bandwidth with regular SATA, you definitely will notice the benefits (especially for async I/O, the dedicated slog device removes a lot of overhead). Secondly: It wouldn't be difficult to implement this as a software solution in windows to be quite honest. Caching models/algorithms are anything but complicated and one could easily write this in userspace or kernelspace. ZFS happens to do this on the file system layer, but it could easily be written on a much higher level of abstraction. There is nothing here that requires an ASIC design level of logic other than maybe freeing some bandwidth on your SATA bus for the bleeding in and out of cache to disk. Thirdly: Microsoft wrote off a feature in vista for moving swap to a dedicated disk (although this was forever possible since windows 2000 simply by specifying a different "page file"). Who's to say that microsoft wouldn't add that to their storage API under the disk management MMC. I can easily see an "add cache device" option being feasibly done.

    2. Re:Good operating system compatibility? by kungfuj35u5 · · Score: 1

      Even further research indicates that microsoft DID do that. ReadyDrive/ReadyBoost can use mulitple devices for on disk caching with their disk prefetching.

    3. Re:Good operating system compatibility? by tepples · · Score: 1

      There is nothing here that requires an ASIC design level of logic other than maybe freeing some bandwidth on your SATA bus for the bleeding in and out of cache to disk.

      Other than perhaps the use case of pulling the drive from one machine and seamlessly putting it in another machine for (say) virus removal or disaster recovery or the like. Without placing the cache inline, the other machine won't be aware of the partially-committed writes sitting on the SSD. But an inline cache will start to commit writes to the HDD as soon as power is restored.

    4. Re:Good operating system compatibility? by kungfuj35u5 · · Score: 1
      Why not wait for the I/O to complete? And in the event of a software implementation, who said that it had to be cached at the block level? You could easily move both if they were NTFS formatted and see both in cache and out of cache contents. This largely relies on the type of implementation, of course, but you never see people using high end raid controllers saying "OH MAN, I CAN'T RISK THERE BEING UNFINISHED I/O ON THE NVRAM!". Either it gets committed or not.

      One disadvantage that some people have sort of eluded to was what happens when this device fails? Is your data obtainable at that point? It certainly isn't in the even that a RAID controller fails (for the majority of them). You have an added loss of data integrity by doing this.

    5. Re:Good operating system compatibility? by tepples · · Score: 1

      Why not wait for the I/O to complete?

      Write-through caching, or not returning from sync() until the data has reached the HDD, is certainly possible. But it would kill the ability to transform (slow) random writes to the file system into (fast) sequential writes to the SSD and commit them in the background.

      One disadvantage that some people have sort of eluded to was what happens when this device fails?

      True, the failure modes of an inline cache would resemble those of a RAID controller. But what happens when the electronics in the HDD fail? As I understand it, it's more likely that the platters will fail first.

    6. Re:Good operating system compatibility? by kungfuj35u5 · · Score: 1

      Do we really want to add to the possibility of failure rates with more electronics in front of a given device? Surely there are much better caching technologies out there than can take advantage of SSDs, in both software and hardware. Who said that it had to be write-through for it to sync to disk fully? Why not just do a full disk sync when powering down? Nobody said that the secondary cache had to be emptied, one can easily detect a dirtied page on the cache. The advantage of write back caching could still easily be felt while the operating system is booted and running. Either way none of these are as elegant a solution as separate volumes, anyway. Who knows how well this thing can even perform when doing the caching, does anybody have any realworld performance data for this?

    7. Re:Good operating system compatibility? by tepples · · Score: 1

      Why not just do a full disk sync when powering down?

      How long would a backup battery in the adapter have to keep spinning the platters after system power down for that to happen? But I guess it could be worked around: flush the cache on power-up, which needs only a working power supply, not a working motherboard.

      Who knows how well this thing can even perform when doing the caching, does anybody have any realworld performance data for this?

      Tom's or [H] might once it comes out.

  22. It seems logical by obarthelemy · · Score: 1

    1- RAM systems work that way too: L1 cache, L2 cache, slow RAM, to compare to RAM cache (OS or controller), SSD, HD.

    2- SSDs right now are very un-optimized: you've got to put, for example, your whole OS on them, even though I'd guess 20-30% of the files are actually read frequently enough to justify being on the SSD... and probably 5-10% of the files are *written* frequently enough to justify NOT being on the SSD. So seeing the SSDs as a cache rather than a hard disk makes a whole lot of sense, and probably doubles or triples their efficiency, by letting them hold only files that best fit the SSD strong points, and hence a lot more of those files.

    My concern is that this "cache"- or "ready-boost"-like mechanism requires quite some intelligence, either, at the most basic level, to keep count of which sectors get read a lot and written not so much, or even, at a higher level, to identify usage patterns and cache the appropriate files (boot, app launch, game play). I'm not sure where that intelligence goes on with the product described... if it goes on at all.

    --
    The Cloud - because you don't care if your apps and data are up in the air.
    1. Re:It seems logical by obarthelemy · · Score: 1

      Another question:

      it seems SSDs are mainly good at boosting boot times, app load times, and game level loading. Why don't HDs load the OS in their cache during POST ? It must be quite simple to memorize which sectors get read first after power-up ?

      --
      The Cloud - because you don't care if your apps and data are up in the air.
    2. Re:It seems logical by ShooterNeo · · Score: 1

      It would require much more complex firmware on the hard drive.

    3. Re:It seems logical by SlashDev · · Score: 1

      "Why don't HDs load the OS in their cache during POST ? It must be quite simple to memorize which sectors get read first after power-up ?" Well that's because POST needs to perform certain critical functions and recognize hardware (such as drives) in order to boot up the computer.

      --

      TOP DSLR Cameras Reviews of the top DSLRs
    4. Re:It seems logical by Anonymous Coward · · Score: 0

      A lot of the time spent in POST is initializing hardware and stuff.

      The only part that gets loaded early in boot is the bootloader and kernel, which are really small. Then the operating system reads in the files needed to boot, usually using a list of files it knows it needs during the boot process.

      Potentially, an OS could copy those boot files to a SSD for faster boot times.

  23. Impressive is that it works with Windows by tepples · · Score: 3, Insightful

    There is nothing new or impressive about this device.

    Other than that it is compatible with applications and peripheral drivers designed to run on the majority operating system for home and office PCs, which has no support for ZFS.

    1. Re:Impressive is that it works with Windows by kungfuj35u5 · · Score: 1

      There is nothing new or impressive about this device.

      Other than that it is compatible with applications and peripheral drivers designed to run on the majority operating system for home and office PCs, which has no support for ZFS.

      Do you not remember the hybrid SSDs that were touted at the beginning of SSDs inception? http://en.wikipedia.org/wiki/Hybrid_drive

    2. Re:Impressive is that it works with Windows by kungfuj35u5 · · Score: 1

      Also, in case you don't read the replies above (which I suspect you haven't because you haven't acknowledged this reply to any extent, ReadyDrive already allowed for this feature in software.

    3. Re:Impressive is that it works with Windows by compro01 · · Score: 1

      I remember them being shouted about with vista's release, being benchmarked with underwhelming results, resulting in the manufacturers playing a blame game with Microsoft over the poor results and who's responsibility it was to write good drivers for the drives, and the devices never actually appearing in stores, followed by the "vista sucks" brigade marching in and the entire concept getting buried as vista was the only thing that supported the drives.

      I'll personally wait for the ATA-8 spec to be released and (hopefully) this implemented in a nice standard manner that plays nice with windows and linux.

      --
      upon the advice of my lawyer, i have no sig at this time
    4. Re:Impressive is that it works with Windows by Ant+P. · · Score: 0, Flamebait

      Fine by me!

      People using a toy OS designed solely for video games and $500 Office Reinstall 2010 can go STFU and keep buying these hardware workarounds for their box of treacle and molasses - after all it's not like they're short of cash. Everyone else will simply have to make do with the steadily growing stream of advances made in filesystem design since 1993.

  24. Breakthrough by thethibs · · Score: 4, Funny

    Good Grief, Alice! They've invented cache!

    --
    I'm a Programmer. That's one level above Software Engineer and one level below Engineer.
  25. 'front-end'? by janap · · Score: 1

    "'front-end' of your hard disk"?

    What does that even mean?

    1. Re:'front-end'? by gimmebeer · · Score: 1

      All files that begin with an "A", obviously.

    2. Re:'front-end'? by Anonymous Coward · · Score: 0

      That would be the purple part.

      Oh, disk...sorry.

    3. Re:'front-end'? by Ant+P. · · Score: 1

      I am guessing it would be the end that's at the front

  26. Re: SSD vs DRAM as cache... by Anonymous Coward · · Score: 0

    Performance of disk systems is determined by the cache system almost entirely (er... given same data transport, etc).

    Usually the write side is more complex because you don't want your device telling the o/s that the write committed when it is still in vulnerable cache. Storage system mfrs recognize this and put in batteries to make sure the write side cache is able to survive power outages. This adds weight as well as complexity and there is always residual risk in these systems.

    If you can make a non-volatile cache write that is effectively equivalent to writing to the magnetic media and do it without external support (eg batteries) then you have a better device.

    Also, this has the potential of bringing high performance writes to the disk device level instead of the storage chassis/system.

  27. zfs hybrid pool by Anonymous Coward · · Score: 0

    Opensolaris can use SSD as cache device for zfs volumes. More details here http://www.filibeto.org/~aduritz/truetrue/solaris10/zfs/zfs-what-next-sdc09.pdf

  28. Re:Shouldn't this be integrated into the controlle by compro01 · · Score: 2, Informative

    Microsoft has a concept they call readydrive for this, mostly for laptops. It was released with vista (Not in XP and I never heard anything about Linux support) and seems to have kinda died. Last I heard anything about hardware was in 2007 with releases from the usual names (Samsung, Seagate, etc.), and I saw a few reviews (which appeared rather underwhelming (supposedly due to poor drivers), which resulted in a blame game between Microsoft and the manufactures over who's fault that was), but I don't think I ever saw the devices for sale.

    There's also plans to include this type of functionality in the ATA-8 spec.

    --
    upon the advice of my lawyer, i have no sig at this time
  29. Re:look at the picture accompanying TFA - it uses by KingMotley · · Score: 1

    I'd rather it not. That way when I upgrade the hard drive, I can toss the old HD and keep the more expensive SSD part.

  30. Makes no sense... by Spazmania · · Score: 1

    From http://www.silverstonetek.com/qa/qa_contents.php?pno=HDDBOOST&area=usa

    After the initial mirroring of data is completed, SSD and HDD will have the same front -end data. HDDBOOST's controller chip will then set data read priority to SSD to take advantage of SSD's much faster read speed. HDDBOOST's priority will be determined by the following rules:

    1.When data is present on both drives, read from SSD.
    2.When data is not present on both drives, read from HDD.
    3.Data will only be written to HDD.

    [...]

    In normal operating system environment, a system drive gets written onto constantly until the system is turned off. Compared to using SSD only as the main system drive, HDDBOOST will only write to SSD once sequentially during system boot up when it activates mirror backup. This significantly reduces the wear and tear that normally occurs when writing data to SSD.

    This makes no sense. How is it supposed to read from the SSD if the SSD doesn't have a current copy of the data because you only wrote it to the hard disk?

    --
    Moderating "-1, Disagree" is simple censorship. Have the guts to post your opinion.
    1. Re:Makes no sense... by obarthelemy · · Score: 1

      my other questions:

      - what is front end data ? data that does not get wiped like my backend's, only shaken ?
      - what algorithm or at least principle do they use to decide what gets on the SSD ? it almost sounds as if they cache the first x sectors, which feels dumb.
      - do I have any control over that ?

      --
      The Cloud - because you don't care if your apps and data are up in the air.
    2. Re:Makes no sense... by Spazmania · · Score: 1

      That part, at least, does make sense. Front end data is the data at the front end of the hard drive. Literally it's the first 1 to N LBA blocks.

      --
      Moderating "-1, Disagree" is simple censorship. Have the guts to post your opinion.
  31. ReadyDrive is not always available by tepples · · Score: 1

    the majority operating system

    ReadyDrive

    As of last month, Windows XP still held two-thirds of market share; Windows Vista and Windows 7 combined made up about one-fourth. Of these, compro01 pointed out that ReadyDrive requires Windows Vista or higher.

    1. Re:ReadyDrive is not always available by kungfuj35u5 · · Score: 1

      While I agree that vista is terrible and I could care less about windows 7, it is still implementable in XP. Even if nobody bothers to implement a utility to do this in software easily for XP (which can definitely be done with the win32 API using simple calls to generate "logical" folders without the need to even touch its driver API), why not put all the fast access stuff on the SSD and use hard disks for the bulky storage?

    2. Re:ReadyDrive is not always available by tepples · · Score: 1

      a utility to do this in software easily for XP

      ...would not necessarily be installed on the machine into which you have transplanted the drive, possibly causing the other machine's view of the file system to be inconsistent.

      why not put all the fast access stuff on the SSD and use hard disks for the bulky storage?

      Ideally, that's what an inline cache like this is supposed to do, except automatically.

  32. Your sig by nuckfuts · · Score: 1

    You know, you don't need a sig that announces "Slashdot reader since 1997". We can all see the number beside you nickname.

    1. Re:Your sig by KlomDark · · Score: 3, Insightful

      // You sound jealous...

    2. Re:Your sig by St.Creed · · Score: 2, Funny

      That's because you might confuse him with other people with low numbers, who only post things and never read anything...

      --
      Therefore, by the (faulty) logic you're using, you're just a cow with a keyboard - osu-neko (2604)
    3. Re:Your sig by nuckfuts · · Score: 1

      // You sound jealous...

      I was merely pointing out a redundancy. I can see your 4-digit number as well, so it's understandable that you'd like to think everyone is jealous. Sorry to disappoint you.

      A low number carries status, but the fact that someone got in early is no guarantee of their sagacity. I judge comments on their content, not on the pedigree of the commenter.

      To put in another way, one doesn't need a low number to post a good comment, one only needs a brain.

  33. Been There, Done That by HTH+NE1 · · Score: 2, Interesting

    Sounds a lot like the CacheCard from SiliconDust for Series1 TiVos, except instead of an SDRAM DIMM it uses an SSD. And the CacheCard doesn't sit between the devices but instead connects to the TiVo motherboard's card-edge connector, provides an Ethernet port, and is designed only to cache a particular 0.5 GiB part of the drive.

    But since the SDRAM loses its contents on power off, it does add significant time to test and fill at startup, while the SSD would be ready nearly immediately.

    --
    Oh, say does that Star-Spangled Banner entwine / The myrtle of Venus with Bacchus's vine?
  34. writes to disk by cenc · · Score: 1

    I am not sure about the speed advantage of reads from disk, given the problem of what to prioritize; but I could see the advantage of writes to disk.

    Does that make any sense?

  35. ... and it's wondows-only by tomhudson · · Score: 1

    You'd be better off just buying a second drive and doing a RAID 1.

    1. Re:... and it's wondows-only by Unequivocal · · Score: 1

      mod up tomhudson. This whole idea is dumb. Buy a second HDD and get raid 1 performance - twice as fast for most activity - not fault tolerant but neither is this solution. This is a new solution to a problem that already has a cheaper more performant solution.

    2. Re:... and it's wondows-only by springbox · · Score: 1

      You mean RAID 0?

    3. Re:... and it's wondows-only by hairyfeet · · Score: 1

      Or hell if you are just wanting the speed boost put a little 32Gb SSD as the OS drive and put your data on the fat HDD. or for those of us on Windows Vista/7 that just want the advantage that SSDs have for random writes just pick up a cheap USB flash and use Readyboost.

      I picked up an 8Gb thumb drive for a whole $20 over Xmas, added the whole drive to Readyboost, and now thanks to Readyboost and hybrid sleep my PC starts up faster than I can reach for the mouse, even after I have had to unplug the power, and thanks to caching on the flash my apps start up crazy fast.

      So I have to agree, this seems to be a solution in search of a problem. You can go RAID 0, you can do as I said and use an SSD for OS and a HDD for storage, you can use Readyboost, there are plenty of options that will give you nice speed and probably cost much less than this thing.

      --
      ACs don't waste your time replying, your posts are never seen by me.
    4. Re:... and it's wondows-only by tomhudson · · Score: 1

      No - RAID 1. RAID 1 keeps identical copies of your data on 2 drives, so whichever one is ready, you can read from ... and if one drive dies, you haven't lost anything. RAID0 isn't RAID - the "R" stands for "redundant", and there's no redundancy built in in "RAID0".. Read a chunk from one drive, then another from the other, then whatever. Since they have the same file layout, the overhead of "keeping track" of where the data is is pretty low.

    5. Re:... and it's wondows-only by nurd68 · · Score: 1

      Last time I checked, Raid 1 was fault tolerant.

    6. Re:... and it's wondows-only by Unequivocal · · Score: 1

      Of course - Raid 0 is what I should have said, as another CP correctly pointed out.

    7. Re:... and it's wondows-only by Vrtigo1 · · Score: 1

      It depends on the RAID driver, some will read from both drives, but I think the Windows driver only reads from one drive.

  36. Why not just make an SSD cache controller? by DutchUncle · · Score: 1

    Why hasn't someone already made this, just like caching IDE controllers? (Terminology wrong, of course, in that SATA doesn't need a "controller" in the IDE sense because it's host-to-host)

    An inline device that you plug in the SATA line. Should be the size of a USB memory stick with a connector at each end, or with an extension cord & connector at one end. Give it 2 to 8 GB of memory, again like a USB memory stick. Different sizes could be different price points.

    Monitor all reads. Cache them while you have empty pages. Obviously the first thing to be read and cached will be the boot sequence on the first powerup after installation, which is probably what you want most anyway, and 2GB is bigger than your core working set even on Windows. On any read, if in cache, return cached copy (obviously), otherwise pass along to disk. Best design will completely avoid delay on the return data by letting it pass through and monitoring it multidrop. Maintain a reference counter on the N pages cached, and on the next favorite N pages (at least); any time a page in cache (or reference count list) is written, invalidate it, replacing the next time you see one of your "next favorite" pages go past. Ditto if a "next favorite" reference count gets higher than the lowest of the N live pages.

    Remember, this thing never instigates action on its own, just piggybacks on system activity. Eventually it stabilizes on your OS, your most-used programs, etc. When you do an update, things get invalidated for a while - your next-frequently-used replace them, until the reference counts go back up. It's self-tuning. The operating system doesn't even know it's there - no driver, no changes, no special code. The disk drive doesn't know it's there. For a frame delay on the SATA request you get acceleration on everything; and if you parse the request in parallel to match you can keep the delay below a full frame.

    1. Re:Why not just make an SSD cache controller? by Anonymous Coward · · Score: 0

      because that's exactly what your OS does anyway?

    2. Re:Why not just make an SSD cache controller? by spire3661 · · Score: 1

      I think the point hes making is why arent we seeing this on the hardware level? A large spinning disk plus a reasonable SSD (5-10 GB) module sounds like a great compromise for the near term future.

      --
      Good-bye
    3. Re:Why not just make an SSD cache controller? by DutchUncle · · Score: 1

      Exactly. The OS does it in expensive RAM with slower access to the platters. Windows ReadyBoost does it with complicated extra software and special hardware and secret knowledge. I want to speed up existing platter drives generically, automatically, and OS-insensitively, with a large percentage of the SSD benefit (one's typical working set) for a low percentage of the cost (because it's a "dongle extension cord" instead of a full SSD).

    4. Re:Why not just make an SSD cache controller? by radish · · Score: 1

      What advantage does that have to the OS level disk cache? Why not just put more system RAM in for that to use than add a special device?

      --

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

    5. Re:Why not just make an SSD cache controller? by DutchUncle · · Score: 1

      It's introducing the low latency of SSD to the *disk drive*, with persistence over powerdown. Plus my motherboard is maxed out. :-)

  37. Re:look at the picture accompanying TFA - it uses by aztracker1 · · Score: 1

    I think that a 32GB SSD in this role would be sufficient for a fairly big boost in performance. I do think that it would be okay to have a 3.5" HDD embed the SSD portion in a similar fashion that this device does, with a 16-32GB SSD and a 16-32MB ram cache on top. As to the more expensive SSD part, well currently pricing has been dropping rather quickly, and would expect to see similar gains for the next couple years.

    --
    Michael J. Ryan - tracker1.info
  38. Already do! rootfs on SSD, home/huge on disk by redelm · · Score: 2, Insightful

    There are alignment tricks with SSD around their large erase blocks, so you have to be careful partitioning.

    Also, consumer-grade MHC SSDs are _not_ tremendously faster than spinning disks in transfer speed. Maybe 20%. Access time is where SSDs shine, 0.2 ms vs 8-10ms .

    A simple scheme I use is to put the OS & small, frequent datafiles on SSD, and large [image] files on platter.

    This might not help large databases with sparse access, but lots of RAM disk cache should be better. IIRC Seagate had a disk with flash boost, but had trouble with it.

    1. Re:Already do! rootfs on SSD, home/huge on disk by owlstead · · Score: 1

      "Also, consumer-grade MHC SSDs are _not_ tremendously faster than spinning disks in transfer speed. Maybe 20%. Access time is where SSDs shine, 0.2 ms vs 8-10ms ."

      My Intel consumer-grade SSD is pushing way more than 20% extra, especially on reads. These things do 250 MB/s, what hard drive gets even close to that?

    2. Re:Already do! rootfs on SSD, home/huge on disk by redelm · · Score: 1

      Yes, I have SSDs spec'd read 200+ . Who knows, they might even do that for 128kB bursts. But when I test them with `hdparm tT` they typically turn _alot_ less, consistently ~50% nominal. 110 is common. Sometimes they run considerably faster on `time md5sum /dev/sda` (140), but never full nominal. YMMV.

  39. Fast stable write caching is the big win by billstewart · · Score: 1

    For a really wide range of applications, what you really need is fast caching of writes to stable storage, so your database transaction logs, file system journals, and modified inodes are saved and the application can go on to its next steps while the SSD box copies itself to higher-latency cheaper rotating machinery. Read-caching is something the operating system can do for itself in RAM (though the SSD box can also do its own prediction, especially for things like track-at-a-time reads of the disk), and the SSD can often keep a bigger read cache than the OS can because it's less performance-critical than caching in system RAM.

    --

    Bill Stewart
    New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
  40. Specifically, Big Cheap Mid-Speed No-Brainer cache by billstewart · · Score: 1

    Yeah, it's a disk cache that's external to your system RAM, with a (hopefully) no-brainer setup. You can use whatever size SSD you want, so it can be bigger than you'll get by expanding your system RAM or replacing your motherboard with one that handles more RAM, and you can use slower RAM or flash for the SSD as opposed to blazing-fast system RAM.

    The big performance win you get from systems like this is write caching on database transaction logs and file system journals, because the writes don't have to wait for rotating machinery to spin around and seek to the right part of the disk, and you can queue stuff on stable storage outside the OS so the application can go on to the next step instead of waiting around for disk interrupts.

    --

    Bill Stewart
    New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
  41. Seagate, WD, others tried this. by CFD339 · · Score: 1

    In 2007 there was a whole movement toward hybrid drives -- it went nowhere.

    --
    The problem with quotes on the internet, is that nobody bothers to check their veracity. -- Abraham Lincoln
  42. Leaping forward into the 1980's by Nefarious+Wheel · · Score: 1

    Haven't disk manufacturers been doing this forever, using faster memories to cache disk?

    Digital's ESE series disks. RAM backed by disk with (iirc) write-behind caching. Expensive (memory was, after all) but in production in the 1980's. Welcome to the future.

    --
    Do not mock my vision of impractical footwear
  43. Windows ReadyBoost by MobyDisk · · Score: 1

    Windows ReadyBoost already does this. Plug-in an SSD, turn it on, and it caches frequently accessed files there. The last benchmark I read on it said it wasn't any faster though - probably because USB flash-drive SSDs are really slow since they are optimized for physical size, data density, and power consumption -- not speed.

  44. Why doesn't by Anonymous Coward · · Score: 0

    Linux do this in software?

    1. Re:Why doesn't by Anonymous Coward · · Score: 0

      1. Partition HD to SSD-sized chunk + rest.
      2. RAID1 first partition and SSD.
      3. JBOD RAID1 and 2nd HD partition.
      4. ???
      5. Profit!

    2. Re:Why doesn't by noname444 · · Score: 1

      I know what ????? is. It's "Pray that your often accessed data ends up on the RAID1-part of your JBOD"

  45. I'd like HDD speeds for my SSD - seriously! by RobWalker · · Score: 1
  46. If it exists, how to set it up? by Ed+Avis · · Score: 1

    OK let's suppose you have a fairly vanilla Linux desktop system, with one spinning hard disk and one SSD. How do you set things up in software to use the SSD as a kind of cache for the hard disk?

    --
    -- Ed Avis ed@membled.com
    1. Re:If it exists, how to set it up? by kungfuj35u5 · · Score: 1

      You setup cachefs in linux.

    2. Re:If it exists, how to set it up? by Ed+Avis · · Score: 1

      Isn't cachefs intended to work with network filesystems like NFS, rather than local disks? I had expected some block-level caching setup would be simpler and give better performance.

      --
      -- Ed Avis ed@membled.com
  47. won't a sata controller do that... by revboden · · Score: 1

    This is kinda like my Vertex SSD boot drive and two 1TB HD raid 1 bulk data setup. Only slower...

  48. dear asshat modbomber by damn_registrars · · Score: 1

    Yest I am posting this reply - to my own reply - just to egg you on. You probably don't have any mod points left or you would have tagged some of my other posts with "flamebait" as well.

    Your attempt to knock me down failed. Indeed it failed miserably as my karma remains at excellent. You really should have found a better way to use your mod points then following every post I made in this discussion with "flamebait". Even better yet would have been for you to actually voice your opposition with a reply, but apparently that was above you.

    In the end, your efforts were for naught. If you ever get mod points again perhaps you will use your points more wisely. Currently I am laughing in your general direction.

    --
    Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
    1. Re:dear asshat modbomber by adolf · · Score: 1

      Wow.

      Listen, dude. I didn't even read your first response to me, after having a peek at your response to yourself and seen just how fucking crazy you are. I hadn't seen either of them until a few minutes ago, since I haven't bothered with seeing if anyone replied to any of my comments for a few days.

      Get a fucking life.