Samsung's 64-GB Solid-State Drive

Anonymous Howard writes "Just a couple of weeks ago Sandisk introduced a 32-GB solid-state drive. Now Samsung has one-upped them, unveiling a 64-GB solid-state drive. They are expecting to begin shipping in the second quarter of this year. Samsung says the device can read 64 MB/s, write 45 MB/s, and uses just 0.5 W when operating (0.1 W when idle). In comparison, an 80-GB 1.8-inch hard drive reads at 15 MB/s, writes at 7 MB/s, and consumes 1.5 W when either operating or idle. No pricing yet."

Comments

finally, one big enough for regular use

[192939495969798999]

64 GB "ought to be enough for anybody"!

Seriously, though, that's enough for windows XP/Vista/etc. plus your favorite games, apps, and so on. Maybe you couldn't put whole slews of videos or images on there, but you could always get 2 of them.

Re:finally, one big enough for regular use

[stratjakt]

You don't need that speed for all your slews of videos and images, just put them ( and all data) on a regular disc, and use this for applications only. It'll last longer that way, anyways.

Re:finally, one big enough for regular use

[Anonymous Coward]

I would end up carrying a firewwire drive around with me for all the other stuff I have. The FW drive would drain the battery, thus defeating half the point of having a SSD in the laptop.

Re:finally, one big enough for regular use

[tritonman]

I would need to buy like 10 of these for my pr0n collection, no thanks!

Re:finally, one big enough for regular use

[PingSpike]

I haven't ever really had any complaints about how long it takes for videos and music to load, which all sit on a little low power file server at my house. However, games and some applications are certainly a different story.

Granted, I probably don't represent your average user. But I do think your average user could benefit from a combination hard drive layout and would even notice the increase in speed!

Re:finally, one big enough for regular use

[nine-times]

Yeah. I could see someone using more than I do, but if I install my OS and all my apps (full installs, clip-art, etc), it's probably going to be under 10GB. Add the games I'm currently running, and it'll be under 20. Add another 20GB for my music collection, and you're still well under the 64GBs listed here. I'll only break 60 GB when you add TV shows, movies, and my software archive (I have a Mac, and whenever I buy software, I make a DMG of the disc and store that for later installs, treat the disc as a backup). 64 GB flash disks would be mighty useful.

Re:finally, one big enough for regular use

[REparsed]

Before these were announced I considered putting XP (lightened) on a 4GB DOM an using a standard drive for data.

Re:finally, one big enough for regular use

[steveo777]

That's actually 4GB larger than my current Notebook HDD. I'm pretty excited to see what the pricing will be.

Re:finally, one big enough for regular use

[Tyler Eaves]

Last longer? The MTBF on flash storage is about an order of magnitude greater than magnetic storage these days.

Re:finally, one big enough for regular use

[MindStalker]

I could be wrong, but I do believe flash turned off stuck in storage has a fairly limited lifetime, while a harddrive has a very long lifetime. Failures from day to day use, yes Flash wins, mainly because small errors can be corrected with regular use.

Re:finally, one big enough for regular use

[FuzzyDaddy]

I could be wrong, but I do believe flash turned off stuck in storage has a fairly limited lifetime

They specify 10 years for flash memory to hold it's data, but in practice (e.g. not at the highest temperature or most extreme operating voltage) it is significantly longer. I don't know to what extent the hard drives work around bad sectors, but they probably do it for both flash drives and the traditional magnetic type.

Re:finally, one big enough for regular use

[nschubach]

On that subject, have you checked out or seen Disktrix's Ultimate Defrag? It relocates the most recent used data to the edge of the platter to apparently lower access time. I'm not sure of any benchmarks against it, but logically it makes sense.

Re:finally, one big enough for regular use

[Solra Bizna]

It's 44GB larger than mine...

...Oh well, at least I'm not on that old 6GB anymore. T_T

-:sigma.SB

Re:finally, one big enough for regular use

[lpcustom]

One Millllllion Dollars!!!!!!

Re:finally, one big enough for regular use

[johneee]

I'm sure there'll be lots of other comments saying much the same thing, but I agree wholeheartedly with you.

I have three laptops in use in my family, and none of them use any more than 30GB on the hard drive... And I could probably clean them up to get them smaller.

Big files get stored on the main desktop at home, and the only things that typically live on the laptops are programs and enough files to keep us going at the time.

If these drives weren't so dang expensive, I'd use them right now and be perfectly happy with the capacity.

Re:finally, one big enough for regular use

[Ramble]

It's enough for the OS but for someone like me who has a TV card 64GB (or 64GiB?) is nowhere near enough.

Re:finally, one big enough for regular use

[compro01]

unless you're planning on treating it as WORM memory, flash has somewhat limited lifespan. the flash we're using at my collage for our integrated processing stuff is rated for 1 million write/erase cycles, but the datasheet doesn't even mention MTBF.

Re:finally, one big enough for regular use

[poadshaw]

No pricing yet.
1 million dollars... mu-haHAHAHAHA

Re:finally, one big enough for regular use

[WheelDweller]

Not room for a slew of videos on 64*G*? Where are you getting these 32G movies, and does the digital production house know you have copies? :>

Most 1-hour videos I'm seeing are 350-700M; there's room for a lot, in there.

What entrigues me is the .5W power use, since I'm looking at vehicular systems these days. I found out last night my 15" CRT is taking more out of the battery than the P3 CPU I'm using! (I knew it was a hoggy, but I didn't know it was that much)

There's always a spun-down 250G raid device you could use, if you wanted to put hundreds of videos in reach- I think the 'lower power mode' can be turned on with hdparm --something, so when they're not in use they're cheaper.

Another thing I like about the solid state drive is the higher reliability and the vibration insensativity; again, useful in a mobile application as well as in laptops.

Re:finally, one big enough for regular use

[Maxo-Texas]

Not sure what it means but 1 of my 3 flash memory cards has gone bad in 2 years.

Looks perfect- but it reports a formatting error when loaded into my camera and reformatting it doesn't fix the problem.

And hard drives last a lot less than they advertise too (all those google related articles 2 months back).

Re:finally, one big enough for regular use

[MyEyesTheyBurn]

Am I the only one that found irony in him going to "collage?" and yet misspelled college? We used some flash + IDE adapter for some wireless routers, and they didn't hold up as we had expected. I believe they were 4GB. Even the with limited Linux logging, they were decommissioned after a year.

Re:finally, one big enough for regular use

[ichigo 2.0]

Assuming a million write/erase cycle limit for this SSD, it would take over 46 years of 24/7 writing to reach the limit on the entire disk. Of course that's unlikely as normal usage patterns would not use the drive in such a way, but I thought it was interesting to know.

(((64 * 1 024) / 45) * 1 000 000) / (60 * 60 * 24 * 365) = 46.1807317

Re:finally, one big enough for regular use

[MrNiceguy_KS]

True, but consider this: These drives will likely be targeted at laptops, particularly the ultra-small, "Road-warrior" type laptops. I know you can get TV cards for laptops, but there is probably very little overlap for TV card owners and users of these drives.

Re:finally, one big enough for regular use

[Anonymous Coward]

Hey hey don't knock its pornographic applications. Just today I gave a friend 11 dvdrs of porn. I'd only need one of these to hold all of that, and it would be significantly faster to transfer the data across to it than it was for the dvds. Finally, when they're done with it, I'm not out anything (drive can be returned and reused, minus write and erase cycles). God bless you samsung!

Re:finally, one big enough for regular use

[triso]

I would need to buy like 10 of these for my pr0n collection, no thanks! The easiest way to fix that is to get married.

Re:finally, one big enough for regular use

[CheapEngineer]

>The easiest way to fix that is to get married.

That'll kill your sex drive enough that you'll have no desire for pr0n again...

CheapEngineer

Re:finally, one big enough for regular use

[hedwards]

Normal use of a flash disk spreads the writes more or less evenly over the disk. The point of this being that a compact flash card that last longer than the first bit which would in all likelihood be rewritten hundreds of times more often than the last couple of bytes on the card.

So yes, it is quite likely that a flash card would be used more or less uniformly over the entire media. The less likely bit is using a card until the entire thing fails rather than just a portion. I would never continue to use a storage medium after it started to fail for anything of consequence.

Re:finally, one big enough for regular use

[binarybum]

darn, I wish you weren't posting as AC. I was going to add you as a friend and hope for the best.

Re:finally, one big enough for regular use

[Phisbut]

Assuming a million write/erase cycle limit for this SSD, it would take over 46 years of 24/7 writing to reach the limit on the entire disk.

As you mentionned, writing on the entire disk all the time is hardly the typical usage pattern. I would be more interested to know if flash drives are becoming suitable for "real" usage, as in, if my swap partition is on a flash drive (swap is always happier with faster transfer), does the drive stand a chance to survive long enough to be useful?

Re:finally, one big enough for regular use

[Phisbut]

Normal use of a flash disk spreads the writes more or less evenly over the disk. The point of this being that a compact flash card that last longer than the first bit which would in all likelihood be rewritten hundreds of times more often than the last couple of bytes on the card.

Typical computer drive usage is far from what we could consider "normal use of a flash drive". Current normal use is usually in digital cameras and other small devices, which users tend to completely erase between use. Therefore, a built-in "data spreader" can easily write data more or less evenly over the disk.

However, if I put such a drive in my computer for regular computer use, there will be one section of the drive which will be erased-and-rewritten fairly rarely (the /usr partition), one section which will be erased-and-rewritten moderately often (the /home partition), and one section which will be erased-and-rewritten quite often (the /swap partition). It would then be hard to spread erase-and-write cycles evenly on the whole disk, unless the drive somehow moves around the big bunch of data which is technically read-only.

Re:finally, one big enough for regular use

[sshir]

I had a similar problem.

Card just appeared broken.

Turned out, I had to re-format it using camera itself, not windows, not an mp3 player.

I have no idea why, but it solved the problem. So I still have fully functional 5 year old cards (all of them).

Re:finally, one big enough for regular use

[eaglesteve]

I got the impression that NAND-based Compact Flash can only support sequential operation, but it faster anb higher capacity than the earlier NOR based flash, the later of which support random access operation. For this reason, the operating systems are usually using the NOR based flash. I believe the 64GB from Samsung is the NAND-based flash, hence not suitable for operating systems.

I'm lazy, yes, but that's not a bad thing

[heinousjay]

I'm wondering, will this work as a drop-in replacement for existing hard drives? The article doesn't say, and while I can't imagine there would be a reason it wouldn't work, I really don't know. In particular, is this something that will work in Vista and not XP?

Re:I'm lazy, yes, but that's not a bad thing

[heinousjay]

Oh, and I read the other article and found my answer. Forget I asked.

Re:I'm lazy, yes, but that's not a bad thing

[gbjbaanb]

yes. a solid-state hard drive is exactly that - a normal HDD that uses a different mechanism for storing data. Usually its a pinning platter, this uses non-volatile memory chips. The interface and size are the same, so you just use it as you otherwise would.

Personally, I think 64Gb is a bit much for me, I'd stick the OS and swap files on there - which come to about 10Gb on my current XP machine.

Re:I'm lazy, yes, but that's not a bad thing

[faloi]

They're marketed as drop in replacements, currently built to notebook drive standards. The downside, as someone else mentions, is that they are flash based. While flash has gotten better recently, I'd be squeemish about having an OS that constantly writes to the drive even when nothing is (apparently) happening on it.

Re:I'm lazy, yes, but that's not a bad thing

[c_woolley]

Yep, that's pretty much what I do with my current system. I have a smaller WD Raptor drive for my OS and important applications, while my RAID drive handles storage and most other applications. Works well. Keeps my system fast and and makes securing it a lot easier when I know exactly what files belong on C:

Re:I'm lazy, yes, but that's not a bad thing

[LighterShadeOfBlack]

There are many reasons why using a swap file is more practical than adding more RAM. For one thing there are inherent limitations on many systems - whether it's a limit on RAM size, a limit to the number of physical slots available, or simply a situation where the rarity of needing additional memory space doesn't justify the cost of adding extra DIMMs. Whatever the reason, "swap files are for memory leaks" is rubbish.

Re:I'm lazy, yes, but that's not a bad thing

[TooMuchToDo]

Don't use swap on the flash disk. Run with a ton of RAM instead. Longer hardware life, with the added bonus of extremely quick data shuffling.

Re:I'm lazy, yes, but that's not a bad thing

[denobug]

Taken from the HD's picture in the article:

1.8" SSD 3.3V ATA7 UDMA66
Model #: MCCOE64G8MPR-03A


Hopefully this helps.

Re:I'm lazy, yes, but that's not a bad thing

[drinkypoo]

One possible reason is an OS that will not run, or will not run correctly, without paging. Linux actually fell into the latter category once upon a time, it would run like crap without swap (unless you made some patches.) SunOS4 was SOLIDLY in that category, but it's old (BSD 4 based - IIRC mostly 4.3, with some 4.4-lite code in the last release of SunOS4?) Windows 2000 would allow you to turn of all paging files, but would blue screen on boot if you did so as it absolutely required some paging (but you could have a teensy, fixed-length paging file.) But in a system that isn't so poorly designed that it requires paging, adding more RAM is a better solution any time you have the slots and the cash. It stops you from paging! That's pure gold right there. Who wants to wait for that?

One solution, of course, is to use a DRAM SSD for your swap if you're out of DIMM slots. Then only cash is the limiting factor. Granted, it still limits me right out of that particular game...

Re:I'm lazy, yes, but that's not a bad thing

[Wicko]

I'm wondering about that also, I mean they have algorithms to spread out data more evenly to increase lifetime, but is that enough to keep it on level with magnetic discs?

Re:I'm lazy, yes, but that's not a bad thing

[misleb]

when I know exactly what files belong on C:

God, i can't wait for the day when we never hear anyone refer to a partition by a letter with a colon after it.

Yeah, off topic, I know.

Re:I'm lazy, yes, but that's not a bad thing

[Thundersnatch]

Your information regarding Windows is somewhat outdated. In my testing, Windows 2000 (with SP4), Windows XP (with SP2) and Windows 2003 SP1 all run fine without any page-file whatsoever. I often run development virutal machines this way to reduce the size of differential backups of the VM file.

When you run without any paging file in Windows, you do lose some crash-dumping functionality, but other than that the Windows kernel doesn't seem to care.

Re:I'm lazy, yes, but that's not a bad thing

[drinkypoo]

I've actually tested it, but maybe I didn't turn off something else related (like crash dumps) or something.

Re:I'm lazy, yes, but that's not a bad thing

[Anonymous Coward]

Where I work we currently have two embedded platforms using 16 gig solid state (flash) drives (laptop form factor). We have Ubuntu 6.10 installed on both. Let me tell you it's very... neat... We have a 1ghz celeron processor, 1gig of ram, and gigabit ethernet on these. All passively cooled and running on 5 volts. If not for the standard atx power supply they were jacked into they'd be perfectly silent.

With this setup ubuntu boots in less than 30 seconds and applications like Open Office load instantly. These are seeing use daily and have yet to fail.

I can't wait for these things to get in the hundred gig range and down near the cost of sata drives. I'd love to get one for my notebook!

Performance?

[26199]

Can anyone find some more details on the transfer rate/seek time?

For a hard disk peak transfer rate is when reading consecutive blocks... if the solid state drive can get near peak performance for random access, it's got a huge advantage.

And is thus very cool.

Re:Performance?

[brunascle]

is there even such thing as seek time with a solid state disk? there isnt really any seeking going on.

Re:Performance?

[stratjakt]

They're called "wait states" and the number of them really depends on the depth of the decoding logic.

You know, those SDRAM timings that x-trememe overclockers 2 the max use to fully run 3dmarks to the xtreme maxx.

Re:Performance?

[Anonymous Coward]

> For a hard disk peak transfer rate is when reading consecutive blocks... if the solid state drive can get near peak performance for random access, it's got a huge advantage.

It should be able to, seeing as addressing a block of flash memory takes orders of magnitude less time than seeking a drive head and waiting for a sector to spin under the read head. Nanoseconds versus milliseconds.

Re:Performance?

[Cemu]

Seek time = 0. There is no head to move and no platter to spin.

Re:Performance?

[Anonymous Coward]

I don't know specifics of this solid state drive by based on my knowledge working with NAND flash I would say the following. For reads, accounting for NAND access time and some command overhead, a 100 us equivalent seek time would be reasonable. Random read performance is probably 90% of sequential read or better. Write performance will be a different story and heavily depends on their mapping algorithms. Without lots of SRAM buffering but a good mapping algorithm, the random write performance is probably 25% of sequential write performance. Buffering will help a lot to improve this.

Re:Performance?

[Anonymous Coward]

There's a benchmarks test for the 32GB Samsung SSD here:

http://www.tomshardware.com/2006/09/20/conventiona l_hard_drive_obsoletism/index.html

The drive's performance seems to be limited only by it's interface (which in this case was ATA/60, not ATA/100 or SATA3). The most telling difference is the I/O results - it achieved around 2000 operations per second, where the mechanical drives peaked at about 50.

Advantage of not having moving parts: huge.

Put pagefile somewhere else?

[Pyrion]

It's flash-based, so am I right in assuming that mapping the pagefile to that drive will dramatically shorten its lifespan?

Re:Put pagefile somewhere else?

[crow]

I'm under the impression that this is solved by two factors.

First, flash parts have internal controlers that remap the flash to level out the writes. (I remember hearing about some researcher who developed a great flash file system, only to find that it didn't make any difference because of the remapping.)

Second, flash parts can handle orders of magnitude more writes now than they could a few years ago.

Re:Put pagefile somewhere else?

[Kadin2048]

Not totally certain, but I would guess so. Depending on how big the pagefile was, perhaps the drive does some intelligent load-balancing that would keep you from frying it too quickly, but it might be better to keep another drive around for that, or loading the machine up with enough RAM to keep it from swapping often.

I used to know people who swore that, after adding RAM, the best thing you could do for speed was to add a small-but-fast SCSI hard drive and use it for nothing but your swapfile. I've never gone that route personally, because I've never thought it worth the expense, but I bet it would make for a pretty nice system. And I also suspect there have to be a lot of SCSI drives on the used market if you look.

Re:Put pagefile somewhere else?

[Anonymous Coward]

> It's flash-based, so am I right in assuming that mapping the pagefile to that drive will dramatically shorten its lifespan?

Finally, a use for all those $5 surplus 7200RPM 8GB drives!

Re:Put pagefile somewhere else?

[Pike]

no, but your browser will need the latest Adobe plugins in order for your operating system to boot properly.

Re:Put pagefile somewhere else?

[Pyrion]

Personally, I'm hoping more for being able to find a Gigabyte iRAM locally now that in a couple of days I'll have about 2GB in 512MB sticks left over after a RAM upgrade.

That's the kind of thing a pagefile should be mapped to - a ramdisk. ;)

Re:Put pagefile somewhere else?

[CastrTroy]

If your pagefile exists in RAM, then why even have a pagefile at all? Just get rid of the pagefile completely. If you have 2Gig of Ram, then you pretty much have no need for a page file for most desktop applications. Things like photo/video editing can take a lot of RAM, but if you don't do any of that, you're probably OK.

Re:Put pagefile somewhere else?

[IDontAgreeWithYou]

Pay close attention to the winky emoticon. It actually means something. ;)

Re:Put pagefile somewhere else?

[Maxhrk]

what?! WHAT? operating system to boot properly rquires adobe plugins?

Re:Put pagefile somewhere else?

[Anonymous Coward]

I think it was the grandparent post's attempt at humor. Since it is flash-based... with flash meaning Shockwave Flash, by Macromedia, now owned by (ominous music) Adobe.

So, if you were basing your system as Flash-Based, you'd constantly need updates from Adobe just to have the OS running.

Of course, that's what I'm interpreting the joke to be. I could be off base.

Re:Put pagefile somewhere else?

[HTH NE1]

Second, flash parts can handle orders of magnitude more writes now than they could a few years ago.

But when it fails, how recoverable is it? Is there an industry for flash memory data recovery like there is for hard drive recovery?

Re:Put pagefile somewhere else?

[Sparr0]

Because some OSes behave VERY differently with and without swap space, regardless of the total amount of RAM+swap. Running linux with ~900MB of RAM and ~100MB of swap (on a ram drive) is far preferable to running with 1GB of RAM and no swap.

Re:Put pagefile somewhere else?

[nschubach]

Are you flirting with me? Police!! This guy keeps winking at me!

Re:Put pagefile somewhere else?

[nschubach]

How does Linux run any different without a Swap partition...and why would it matter?

Re:Put pagefile somewhere else?

[drinkypoo]

Frankly it doesn't even matter if you use SCSI or not. What matters is that you're not using a controller that offloads all processing to the CPU. Some SCSI controllers even did this, so they won't provide any speed bump over using onboard IDE. Most SCSI controllers DO have their own processor though (amusingly, even an Adaptec 2940U has a Z80 chip on it, it's the chip that would not die!) and most ATAPI controllers don't. Any HW RAID controller does, of course. Any SW RAID controller probably does not although they typically have enough adapter BIOS to get the system booted... but that code runs on your CPU. ATAPI and SATA controllers that are non-raid sometimes have their own microcontroller, sometimes not. USB is actually the same in this regard but for some reason even the kind with a coprocessor seems to use more CPU time than just using IEEE1394. SCSI does have one advantage over most other drives, which is tagged queuing. Without going into it, queuing reduces the number of seeks, which as you can imagine is pretty damned handy when it comes to swap. SATA does it too, and some high-end ATAPI devices as well IIRC (but you need UDMA4 or something.)

Re:Put pagefile somewhere else?

[benzapp]

Flash drives are designed to operate with the same level of reliability as mechanical hard drives during the warranty period. After the maximum writes is reached, there will be irrevocable data loss - but that will occur after the warranty period, which is at least 3 years.

Only a fool runs anything of importance, without a backup, on a hard drive that is over 3 years old.

So the answer to your question is simple: Who cares? For the near future, flash drives are marketed towards people with money who want a lower power, fast hard drive. It is not marketed towards cheap geeks who want to keep their hard drive for a decade.

Re:Put pagefile somewhere else?

[Pyrion]

Gigabyte iRAM
It's technically a ramdisk because it's storing data in RAM sticks, yet the RAM is seen by the system as a SATA drive.

Re:Put pagefile somewhere else?

[asavage]

Fortunately flash fails on writing not on reading so while you can't write to sectors that have failed you can still read them.

Re:Put pagefile somewhere else?

[egomaniac]

Only a fool runs anything of importance, without a backup, on a hard drive that is over 3 years old.

That sentence should have ended right after "without a backup".

Re:Put pagefile somewhere else?

[Overzeetop]

Would you people just give it up on the read/write FUD?

Modern NAND is good for 1E5-1E6 cycles, and then some. Go figure out how long it would take - assuming write levelling - to take one of these discs to their limit, given a 3% pagefile and 50% typical capacity utilization (without dynamic reallocation). I suspect you'll find somewhere in the answer given in years of continuous uptime.

Here...I'll do it, with rounding for ease:

32 GB space for swap file - say 30GB to save one chunk for safety. If we presume one read per write, we get about (64+45/2) 55MB/sec, half of which is write, so 28MB/s.

Drumroll please...32e9*1e6 / 28e6 s^-1 = 36.2 years if you do nothing but write-read to the disk swap and do nothing else (sounds like Vista, huh?).

I've got this funny feeling you won't run out of cycles on this drive.

(btw - with dynamic reallocation of space - writing existing locations with low cycles and low turnover to areas with higher cycles, one could use the entire drive space. This is possible since the random access is nearly the same speed as sequential access - the pagefile does not need to be a contiguous physical block of memory)

Re:Put pagefile somewhere else?

[Mr Z]

Except in the field of computer forensics, is there a market for page/swap file recovery?

Re:Put pagefile somewhere else?

[Mr Z]

The numbers you used, approx 900/100, are also a special magic point on 32-bit CPUs under Linux. Above about 960MB, Linux uses "highmem" mode on x86, and that slows things down. A 32-bit x86 PC runs faster when you restrict it to 960MB instead of letting it use the full 1024MB.

For those of you who wonder how a computer could run faster w/ a little swap in RAM instead of just using all the RAM, the answer is complicated. Mainly, all the VM algorithms assume the existance of swap, and so when they get backed into a corner, they expect to be able to dump a bunch of stuff overboard into swap. They only start making the really hard choices once swap fills up. If you take away swap, then you hit the "out of swap" condition much more readily.

You might be thinking "ah, but it's all just a shell game! You'll still run out of swap at the same time, since your total memory is fixed!" Not true. The OS prioritizes disk buffers and other caches relative to the work it's doing and the RAM available to it. RAM dedicated to a RAM disk is not available for other purposes. Thus, a RAM-based swap partition dedicates some portion of RAM to only hold dirty program pages. No disk buffers, no network buffers, no inode information. Just dirty program pages. By forcing austerity on these other discretionary structures, you can compensate for the VM's inbuilt assumption it can just "dump things to swap", and that running out of swap occurs "almost never."

--Joe

Re:Put pagefile somewhere else?

[KozmoStevnNaut]

And it's only a viable product due to the ridiculous design of a certain OS.

If only swap was handled properly on Windows, the iRAM would never have existed, since you would just add more RAM to your machine and switch off swap completely, which you can on a proper OS.

Re:Put pagefile somewhere else?

[MrNaz]

Even if there is, would it make a difference? The industry that exists for data recovery only exists because people don't know to backup. Drive failure is inevitable, there should never be reliance on any single device, whether it is 10 minutes or 10 years old.

What's the long-term stability?

[davidwr]

Quality hard drives are fairly reliable. They can last 10 years or more and you can usually count on them to last their warranty period - 3-5 years - and then some.

They also have error detection/correction, bad-sector remapping, and "I'm about to die" notification.

At one time, solid-state devices were good for about a thousand writes for any given memory cell, a lot fewer than HDs.

Does anyone know the reliability for these new solid-state devices over wall time, hours in use/plugged in, number of read cycles, and number of write cycles under normal operating conditions, and how those compare with a modern 1.8, 2.5, or 3.5" drive?

Re:What's the long-term stability?

[Timoteo47]

The SSDs from Samsung and SanDisk will last for years and have an MTBF of 2 million hours. San Disk claims there device will last at least 5 years.

Re:What's the long-term stability?

[Chacham]

"I'm about to die" notification.

Ah, the old IATD. Known to the rest of the world as Innovation Alliteration for Termination Declaration. Had they only known.

Re:What's the long-term stability?

[MrZaius]

For comparison's sake, apparently some of my 250GB WD hard drives have a MTBF of only 1 million hours.

Re:What's the long-term stability?

[Lord Ender]

Boy is your information wrong! Do you sell hard drives, or do you get all your information from drive salespeople? Haven't you read the intertubes in the past few months?

There have been TWO scientific research papers published recently which demonstrated that the likelihood of a drive failing increases linearly with time. You CAN NOT COUNT ON THEM to last their warranty period. These studies also demonstrated that SMART (the "I'm about to die") signals are only useful for a minority of possible drive failures. You can't count on that, either.

Additionally, I've had servers running off of solid-state drives for over a year. It's great. I did have one solid-state drive fail in that time--but NOT due to write-limits. I believe it was a power issue.

Don't spread misinformation.

Re:What's the long-term stability?

[emcmanus]

I was scanning the comments looking for someone discussing solid state drives (whoops, almost said discs) with server applications.

Any idea on the kind of performance increase you've seen from making the switch?

Re:What's the long-term stability?

[volsung]

The recent Google paper about disk failures that pointed out that SMART is not as reliable an indicator of impending disk death as you would hope. One third of the failed disks had no SMART errors reported at all. (Two thirds is not bad, of course, but you really want something a bit more deterministic.)

Re:What's the long-term stability?

[Tom Womack]

I believe the current figure is something more like a hundred thousand writes per memory cell; with a load-distributing file system, this means the lifetime is about the amount of time taken to write the whole device 100k times, which (since flash writes are fairly slow; a few megabytes per second at best) is up there with hard-drive lifetimes.

Flash can easily be made to have ECC and bad-sector remapping ... if you look at the spec sheet http://www.samsung.com/Products/Semiconductor/NAND Flash/SLC_LargeBlock/16Gbit/K9WAG08U1A/ds_k9xxg08u xa_rev11.pdf (linked to from an HTML summary at http://www.samsung.com/products/semiconductor/NAND Flash/SLC_LargeBlock/16Gbit/K9WAG08U1A/K9WAG08U1A. htm ) for a Samsung flash chip which is probably the one in these drives, you'll observe that it has 256Mbits of spare capacity for that sort of purpose. Doing reads and writes in units of a whole sector means you can use quite efficient error-correcting codes.

Re:What's the long-term stability?

[icepick72]

To lend credibility to your post, please cite the "TWO" (in capital letters) scientific research papers you are referring to. Also, point out the sections which prove that you "CAN NOT COUNT ON THEM" (again your caps). BTW, Caps don't make your post more accurate for lack of links to proper information.

If you accuse grandparent of getting info from wrong places, then why do you not provide sources to back up your claims? (Do as I say, not as I do?)After you provide the info we can analyze it. TIA

Re:What's the long-term stability?

[Lord Ender]

Considering that both papers (one published by google and one by a CMU researcher) made it to the front page of slashdot, I'm sure the minority of readers who missed them the first time could look them up very easily. I don't care about the issue enough to do their searching for them.

Re:What's the long-term stability?

[Lord Ender]

Well, my systems are not switched over from something else--they are new systems. I needed several multi-TB arrays and I wanted them to be highly flexible (so Linux software RAID, not hardware RAID), and low maintenance so that I don't have to fly around the freakin world if something breaks in India or Tokyo (I don't want their bootup to be dependent on a mechanical device working).

I ended up getting these tiny little 4GiB flash drives that clip right into the IDE ports on the motherboards (no cables or screws needed). I configured Linux (Ubuntu LTS) to reduce writes to the flash drive (by mounting tmp and var on the raid, and mounting root noatime, etc.). This way, I can always log in and see what's wrong, no matter how many drives fail. And since it's software RAID, I can resize (or whatever) to fix the problem remotely without cracking the case open.

They actually have an order of magnitude lower sustained reads (5M/s compared to 50M/s for SATA) but their seek time is, of course, nearly 0. Since sustained reads are only done from the flash disk during bootup, there is a significant increase in overall performance.

It's a great system, and designing it was a lot more fun than most of the IT maintenance crapwork I do every day.

It cracks me up that people still spend x-bajillion-dollars on SAN equipment from IBM when you can do the same thing for a few thousand by taking advantage of the newest tech.

Re:What's the long-term stability?

[icepick72]

I don't care about the issue enough to do their searching for them.

It's obvious because you still haven't been able to classify your argument "You CAN NOT COUNT ON THEM to last their warranty period". Yet you used bold and caps to emphasize your argument so how unimportant is it to you really? Your argument is not to be considered valid unless you take serious onus to support it. Naming a couple of papers does not ensure they contain the information to support your argument.

Re:What's the long-term stability?

[Lord Ender]

Well, the data is out there. Engineers who know it exists will take it into consideration when designing their systems. Engineers who don't will continue to design inefficient (too much redundancy) or risky (too little redundancy) into their systems. I'm not going to do it for them, but I didn't mind pointing them in the right direction.

Re:What's the long-term stability?

[Plekto]

Nice summary.

What's the biggest problem with hard drives? Right - heat. You need fans and all sorts of wizardry for a typical rack/array. Imagine these dropping to zero noise, zero heat, and 1/3 the power used.(minimal at best).

Suddenly you don't need that $100K forced cooling system to keep the room cool enough so that they drives aren't sucking in 100+ degree air and melting themselves in the span of a month or two. If the fans or cooling dies... yep, the system goes into shutdown mode in minutes. Seen it happen many times. It's not the energy useage for the drives that's the big deal, or the fans, but the fact that there's so much waste heat in a typical array that you need external cooling of the room.

Now imagine no 10 ton or larger A/C system. No holes in the building, no engineering. Just roll the flash-array (smaller, too) in the corner and plug it in. 200 flash drives in a brick the size of a small rack.(virtually no heat means you can stack them like legos) At most, a big heatsink on the rear would suffice.

And it halves the typical heat load for a blade or even a desktop. There's no 150+ degree drive in the case needing loads of airflow(most are almost painful to the touch they run so hot). Longer system life as well.

This will obviously be the future. All we need is a bit more reliability/lifespan and a lower cost. I'd say no more than 2-3 years before we all are seeing flash drives in desktop computers. Maybe 1-2 years for laptops.

Q: how reliable are those boxes? 4gig is low, but 16gig would work for a Ubuntuu install and leave plenty of room for apps. Have you tried this with a raid aray? "Oh gosh - my 4 gig CF stick blew up after 3-4 months"(plugs another in for $20). :)
(I'd suspect that the drives would be hot-swappable as well?)

P.S. the spellchecker here doesn't recognize "Ubuntuu" - lol.

Perfect for MP3 players

[Grishnakh]

This would be perfect for my iRiver H320 MP3 player, since (according to TFA) it's in the 1.8" form factor which almost every HD MP3 player uses.

Re:Perfect for MP3 players

[DeadChobi]

I'm wondering how much disk my Zen Sleek Photo can use. If I can get a 1.8" 64GB flash drive and replace the solid state drive with it then it'll have better uptime than my cellphone. All I need is a better battery and I'm golden. The question is whether or not it puts out more or less heat than the hard disk. BTW, there's probably a howto for disassembling your MP3 player somewhere that you could use. The remaining question is whether or not the firmware will support it.

Re:Perfect for MP3 players

[maxume]

It uses less energy. It will put out less heat.

Re:Perfect for MP3 players

[Anonymous Coward]

Perfect drive indeed. As long as you don't mind you mp3 player costing an extra couple grand ;)

Not that I'd ever buy anything Apple makes, but they make one with a 60GB HD if you need that much space. I doubt the power saving would make a huge difference in battery life either.

Personally, my current 40GB model has enough space as it is (even to put some photos and documents on it when I travel), and the battery life is good enough that I only have to recharge it once a week or so. For 50$ more, maybe I'd get this, but it's not worth the price they'll be asking for this.

Re:Perfect for MP3 players

[Grishnakh]

I don't know about your Zen, but my iRiver is known to support 80GB drives with the stock firmware, even though the device never came from the factory with more than 40GB. If your device came in models with different HD sizes, it's likely it just auto-detects the size and uses it, though there's probably some upper limit.

Better battery: check out newertech and other iPod battery retailers. If your battery is close to the size of one of the iPod batteries (remember, there's a lot of sizes with all the generations and lines (regular, nano, mini, etc.)), you can probably fit one of those in there with a little wiring work.

Heat: as someone else said, if the drive uses less power, it creates less heat.

The big questions for all of this are: will these 1.8" flash drives be available to the public, and how much will they cost?

Re:Perfect for MP3 players

[Grishnakh]

I'd be happy, for right now at least, with the 32GB model. The power savings would make a huge difference in battery life, however, since the HD is what consumes most of the power in an MP3 player. If these use half the power or less, that's probably at least a 33% increase in battery life (the electronics use some too). That equates to several hours.

You're right though: the prices these sell for will probably be too much. Not thousands (or else they'd never be considered seriously for laptops), but still too much for an old MP3 player. Maybe we'll get them used on Ebay after a while.

I don't get it...

[bluemonq]

How can it be one-upping them A-DATA already annouced 128GB SSDs two months ago?

Re:I don't get it...

[brunascle]

the samsung is a 1.8 inch drive and the A-DATA is a 2.5 inch.

Re:I don't get it...

[Klaruz]

If you read the actual source. Instead of the blog post you'd see they have a 64gb 1.8" IDE version too. I'm not sure if the current interface standard for 1.8" drives is IDE or S-ATA, but I'm sure they'll put the appropriate connector and interface chip on whatever they sell when they ship them.

Re:I don't get it...

[nschubach]

It might be a speed factor. As far as I can tell, no stats were dropped on the A-Data drives besides size and format.

Re:I don't get it...

[Kanasta]

My ADATA 150x speed SD card runs benchmarked 0.8MB/s, slower than my Panasonic 10x speed SD card at 1.2MB/s.

I wonder if this Samsung speed relates at all to the HD speeds? Cuz I know HDs REALLY can read 15MB/s in REAL LIFE on FORMATTED disks. No flash tech so far has ever published real life speed yet.

Re:I don't get it...

[A_Non_Moose]

the samsung is a 1.8 inch drive and the A-DATA is a 2.5 inch.

IOW, "Mine's smaller than yours...ha ha!"

Wait, I didn't mean it like that! Stop laughing!

From my vantage point

[rolfwind]

Hard drive capacity growth has slowed the last years in notebooks, they just haven't been increasing in size that fast as in the early 00s. I think flash will surpass notebook harddrives in size within 2-3 years. As it is, 64GB is in the same magnitude of existing typical notebook drives now, just halfway down on the scale.

The price may or not go down enough within that time period to kick out harddrives completely - in which case we'll just see hybrid drives take over.

Re:From my vantage point

[timeOday]

Hard drive progress has dramatically slacked off in general, not just in laptops. here is my recent usenet rant on the topic. The upshot is that if trends from 2001 had continued, you could now buy a 5 terabyte drive for $300. Instead it's $300 for 750GB.

Re:From my vantage point

[typicallyterrific]

Actually, I'm not so sure it's so bad.
I mean, you're not always gonna get returns of 50% more memory every two years, but on the other hand you *did* just gain an extra 50 gigs.

Iunno about you, but that's a still a very usable improvement curve.

Seek time? Should be 0ms

[CFD339]

There should be no seek time, it's solid state. There is no read write head to move, and there is no platter to spin.

Smaller isn't better...

[Channard]

All this means is that we'll suddenly get smaller and smaller MP3 players. Which I'm not sure is the right direction. I'd rather have the players stay the same size and for the batteries to get bigger. I'd happily trade miniscule size for a much longer battery life.

Re:Smaller isn't better...

[Pyrion]

It's flash-based, so I would think the energy savings from not having to constantly run a hard drive's motor would lengthen battery life just with the batteries as they are now.

Re:Smaller isn't better...

[eebra82]

I'm not sure what you're getting at here. Just because manufacturers can create a GSM phone the size of a small watch doesn't mean they are doing it. Some niche products will always go beyond limits but there is a practical limitation as to how small a gadget can get, and the developers know this.

Re:Smaller isn't better...

[gmb61]

I'm not so sure they do. Have you ever seen a MicroSD card? They're about half the size of your thumbnail. I'd be afraid to eject it from my device for fear of dropping it and never finding it again.

Usable life...

[Cervantes]

I'd love to grab one of these (even the 32MB one) and slap my OS and apps on there... but I'm concerned about what the usable life really is. I mean, sure, it's good for maybe a million write cycles (number pulled out of my ass), but really, with your OS running, the usual memory-resident programs, perhaps a nice game of warcrack going... how long is that going to last you?

Some systems I have use hard drives I bought ten years ago... really, 8GB is more than enough to hold the OS, programs, etc etc, and if it's working and I have it backed up, why should I bother buying a new one for something that doesn't need it? Will these drives hold up long enough to be used 5 years from now? How well do they degrade? Do you just notice that your available size gets smaller over time, as the flash gradually goes bad and the bits are eliminated from the FAT? Or does your system stop working one day 6 months from now because the bad bit was in your bootloader?

I'd love to get my hands on one of these, HD bottlenecks are the biggest PITA.... but I'm not going to be an early adopter. I want to hear some horror stories first.

Re:Usable life...

[Fizzl]

I'm under the impression, that solid state drives cycle through the available space somehow, so all the blocks would get approximately equal read/write cycles. So that a) it lasts as long as possible b) when it starts having bad blocks, you better replace the thing pretty fast.

Re:Usable life...

[LWATCDR]

So go for it.
http://www.psism.com/adcf.htm
Has CompactFlash to IDE adaptors.
Get a few of these and some Compact Flashcards and then set up a Flash based raid. I Would keep a my swap on a regular drive but modern motherboards tend to have a few IDE slots and a few Sata connectors.
Could be a cool little system.

Re:Usable life...

[gwern]

I'm going to interpret your first comment as meaning, "It'd be neat if I could mount the / partition on the solidstate drive, and /home on some sort of regular disk-based harddrive." In which case, why not? How often do your configuration files in /etc, or kernels in /boot, or your binaries in /usr/bin, or files in /root change? Not very often, I would expect, even if you are upgrading constantly - how many apt-get upgrades or apt-get installs would it take before sectors start going bad and reducing space? A lot, I'd figure.

(Now, if you want to argue that something like /var or /tmp should be in RAM or on a diskdrive, that'd be sensible, but you could still have the vast majority of your system stuff in the solidstate drive.)

Re:Usable life...

[benzapp]

Flash Hard Drives are designed to work within the warranty period. Like modern hard drives, there is a limited number of writes available, and the actual capacity of the drives is greater than the accessible capacity. Your 320 gig hard drive is really more like 400 gigs, and the 64 gig flash drive is more like 80 gigs. The drive intelligently spreads the writes around as there is no seek time, but reports the usable amount to you.

Plan on using the drive for 3 years and you'll be fine. If you want to use a drive longer, which in and of itself is probably not very wise, get another kind of drive.

Price: $200ish?

[crow]

Based on 4GB compact flash prices at Pricewatch, I can get 32G for $107.60 or 64G for $215.20. All that's new here is packaging all that in one package, and putting a regular IDE interface on it. So at today's prices, that's about $200 per 64GB drive. Of course, by the time this hits the market, it should be lower. On the other hand, there will be a significant premium charged at first until there's enough competition to bring it down.

Re:Price: $200ish?

[Phroon]

Based on 4GB compact flash prices at Pricewatch, I can get 32G for $107.60 or 64G for $215.20.

No, you can't. The $13.45 == ($107.60/(32GB/4GB)) == ($215.20/(64GB/4GB)) 4GB CF card your looking at is out of stock. (Possibly because you mentioned it here...)
The next cheapest (in terms of $/GB) is a 4GB Pen Drive for $32.95. That leads to prices of $263.60 for 32GB, and $527.20 for 64GB. Much closer to what I'd expect for flash memory.

The only reason I checked is because I wanted a $13.45 4GB CF card.

Re:Price: $200ish?

[Anonymous Coward]

WATCH OUT. The company that is selling this just jacked up the price about $30. Buyer Beware.

Re:Price: $200ish?

[tygt]

I've got a little portable media player/flash reader/disk drive unit that I got for storing digipics while on the road. It eats batteries something fierce; perhaps I'll swap the HDD for one of these flash units and it'll last a bit longer.

I would spend serious money for a laptop drive

[CFD339]

What would you spend if you could be a 2.5" version that was interface compatible with your laptop sata connector that was say, 100gb with comparable power and performance?

Personally, to pull the SATA drive out of my laptop and replace it with a 100gb version of this that used so much less power and was so much faster would be a no-brainer even at something like 700 or 800 dollars (US). Battery life would be radically better, noise and heat would be much lower, performance better and general usability should be outstanding.

What are the downsides? How is the duty cycle on these things? Will they last as long or develop hotspots that can't store data as well?

Re:I would spend serious money for a laptop drive

[zippthorne]

one more downside, perhaps: zero noise. How will you know the difference between a hang and simple thrashing without being able to put your ear to the chassis and listen to the disk doing what it does?

Re:I would spend serious money for a laptop drive

[ambrosen]

Well, you can get 32GB for that price right now. Although I've only seen it sold as PATA, for reasons I do not know.

Re:I would spend serious money for a laptop drive

[ivan256]

When my hard drive is thrashing, the access light comes on... You should look into hooking up that cable. ;)

Re:I would spend serious money for a laptop drive

[zippthorne]

Mine, too. But when my system hangs, the light *stays* on, even if the hard drive isn't doing anything at all.

Re:I would spend serious money for a laptop drive

[cerberusss]

Battery life would be radically better

It would be better, but not radically so. Looking at these numbers and generalizing a bit, it would mean that you would get a 25 percent gain in battery usage.

Re:I would spend serious money for a laptop drive

[ivan256]

Lame. That shouldn't happen with SATA 2 drives, because the activity signal comes from the drive itself. SATA 1 drives and older P-ATA drives could be an issue though, if the controller decided it should implement the light via software.

Re:I would spend serious money for a laptop drive

[frdmfghtr]

When my hard drive is thrashing, the access light comes on... You should look into hooking up that cable. ;)

Dammit I use a MacBook and don't have a drive light, you insensitive clod!

(I couldn't stop myself...sorry)

Re:I would spend serious money for a laptop drive

[Jeff DeMaagd]

If you are using a standard notebook, then I think the extra battery live you'll get with a flash drive is about ten minutes. You'll only get significant extra life if your computer is an ultraportable with ULV type chips.

I haven't had a noisy or not notebook hard drive, though I haven't owned notebooks for long. I have a 1 yr old notebook and one that's over 4yrs old and I can barely hear them when they are seeking if I put my ear close to the drive. Usually it's the old drive types that are loud. If it's loud and hot, then you'd be better off bitching to warranty support or just spending $100 for a replacement drive if it is out of warranty. I'm certainly not going to spend $800 on a drive when I can get 160 or 200GB for $200.

MTBF

[EssTiDee]

The SanDisk 32GB version reports a 2 million hour MTBF... http://www.sandisk.com/Oem/Default.aspx?CatID=1478

That's quite a bit better than typical hard drives these days!
Has anyone found MTBF information regarding the Samsung versions?

Re:MTBF

[Anonymous Coward]

MTBF can be really deceptive - it's not about wear-and-tear as most people think but about general reliability. In particular, int this case it doesn't take into account the limited number of write cycles permitted by flash memories (usually about a 1 million per cell, much less than a harddrive which is worn in a completely different way).

Longer story: A MTBF of 1.000.000 years doesn't mean that if you take a sample of units and let them run until they've all failed, the mean time before failure (averaged over all units in your sample) will be 1.000.000 years. What it rather means is that if you take 1.000.000 units and start them at the same time, then you would expect the first unit to fail within 1 hour.

Heat and Noise?

[lawpoop]

I read the articles. I didn't see anything about heat and noise output. Can anyone fill me in? I would guess it would be minimal and none, respectively.

Re:Heat and Noise?

[amorsen]

I read the articles. I didn't see anything about heat and noise output. Can anyone fill me in? I would guess it would be minimal and none, respectively.

Well, based on an energy consumption of 0.5W and an educated guess that they probably aren't emitting much light, I'd say that the heat output is 0.5W.

Duh.

Re:Heat and Noise?

[njchick]

... which leaves 0W for the noise.

Re:Heat and Noise?

[Anonymous Coward]

...because electrons are known for screaming at the top of their tiny, tiny lungs as they course through solid state circuitry.

Re:Heat and Noise?

[Anonymous Coward]

Emit light? 0.5W of heat? these arent light bulbs.
Since when do you measure heat with watt?

Re:Heat and Noise?

[Anonymous Coward]

Since thermodynamics was invented. http://en.wikipedia.org/wiki/Heat

Re:Heat and Noise?

[linuxmop]

You're right. We should be using joules per second.

Re:Heat and Noise?

[syncrotic]

Actually, even if it were emitting light, that would still count as heat. Visible light is no different from infrared: when the photons are finally absorbed by something, the temperature of the absorbing material increases.

An incandescent lightbulb, for example, emits about 95% of its energy in the infrared part of the spectrum and 5% in the visible part. Both infrared and visible light, after bouncing around for a while, eventually get absorbed by the walls of your house, and those walls will heat up slightly.

Re:Heat and Noise?

[amorsen]

Sure, light mostly ends up as heat too, but you can't be entirely sure that the light doesn't go through a window and maybe, just maybe, ends up going into space. Fairly theoretical for a hard drive in an enclosure.

I'd like one in my next iPod

[soft_guy]

I didn't RTFA, so I don't know how physically large these are, but I want one for my next iPod. It isn't a high enough capacity for a laptop HD for me, though.

Re:I'd like one in my next iPod

[wickedsteve]

The physical size is iPod size, 1.8" But there are actually thin 1.8" drives and thick drives. That is why the 30GB iPod is slightly thinner than the 80GB iPod. Thicker drives have more platters. I would happily trade my 80 GB HDD iPod for a 64 GB SSD iPod just for the advantages SSD provides even if it is less GB. My music library isn't even 30GB anyway.

Industrial PC's

[timias1]

I worked with some industrial PCs back in 98, and they came equipped with solid state hard drives. They were around 120 MB, but the could load Win 98 in a few seconds. They solid state technology was important in that application, because it was highly resistant to shock and vibration. They could withstand like 80 g's of shock. Is there any reason that solid state cannot ultimately replace the current HD technology? It seems like a logic progression. Horses to Automobiles Propellers to Jets Vacuum Tubes to Transistors.

Re:Industrial PC's

[stratjakt]

Is there any reason that solid state cannot ultimately replace the current HD technology?

$$$

For a long while I think you'll see more hybrids, and more use of a solid state drive to accelerate application loading, while platter based discs hold the mountains of "data".

Other than application loading, there isn't too much use for these on personal PCs. They'd improve the hell out of database server performance, though.

Re:Industrial PC's

[eebra82]

Of course solid state disks will replace regular hard drives. After all, the conventional disk is the only computer peripheral with moving parts.

I think that the SSD is going to compete far sooner than most people realize. Looking at the numbers, we now see that laptops are almost outselling stationary computers, so people may actually turn to SSD as soon as 2.5 inchers at 200 GB come at competitive prices. Besides, if you want lots of space for vids and mp3s, then why not get a networked server with a couple of TB of space, or at least some external drives mounted to a laptop slot-in?

Re:Industrial PC's

[RiotXIX]

as pointed out, $$$.

But do you really need ssd in large volumes?
No os partition + apps should require more than 10GB, and that gives you about 20GB+ of temp data to play with.

For my next setup (finally building THE dream machine), I'll have a 32gb (actually would have been content with 16gb, but I guess I can rip more large wavs/mov's in one go), I'll have my main parition/documents on the flash drive, and then a big SATAII mechanical drive for storing this large, infrequently accessed files on. I'd have it no other way, even if larger flashes did exist.

I actually figured it would be a nice idea to have a seperate tiny drive for my OS/ACTUAL work, and another for media. This will be convenient for such a setup. And it will be very easy/clear to seperate the wheat from the chaff when backing up.

Re:Industrial PC's

[garyok]

After all, the conventional disk is the only computer peripheral with moving parts.

Ummm... CDs and DVDs (not to mention the blinking-flip 1.4Mb floppy drive I still need to load RAID drivers on XP)? CPU and PSU fans? My printer? My (opti) mouse's buttons? The front door thingy on my PC case? Still lots of moving bits around in conventional PC peripherals that can wear out, 'fraid to say.

Re:Industrial PC's

[repvik]

Of course solid state disks will replace regular hard drives. After all, the conventional disk is the only computer peripheral with moving parts.

Well, I guess your CPU fan, PSU fan, CD-ROM and possibly floppy drive is broken since they're not moving?

Re:Industrial PC's

[Anonymous Coward]

You forgot the main peripheral with moving parts that tend to wear out... the user.

Re:Industrial PC's

[Wicko]

While the advantages are clear, regular hard drives won't disappear. They are much cheaper for higher capacity, as they keep releasing huge drives (Toshiba had a Terrabyte drive?). It will always be good for such things like servers and back up data that you don't need to store on flash drives (like music, movies, etc). Plus, hyrid drives are coming if they haven't already, and that adds a lot of performance for much cheaper than solely flash based, in addition to the large capacities. Magnetic disc drives aren't going anywhere anytime soon.

Re:Industrial PC's

[eebra82]

Have to disagree with your thinking. I was obviously thinking of the most vital parts of the computer. I agree that the CD/DVD player could be included in that list, but a computer fan is only optional. You can get coolers with no moving parts too, you know.

I am not sure if you're being sarcastic, but I was talking about the computer and not any mice, printers or whatever.

Re:Industrial PC's

[Kjella]

But of those mentioned, the only one that actually stores the data (CDs/DVDs store it, but the drive doesn't) of those mentioned. The others can be replaced, failed hard disks are usually a Emergency(tm). Yes, there's RAID and live/nearline/offline/offsite backup. No, people still won't do it. From what I've understood the SSD disks will be more reliable. They have a limited lifespan but it should be more predictable. A HDD might be a microscopic flaw in its bearings or motor or disk heads, which after a year of spinning at 7200rpm makes it crash and burn. Unless it has shorted out for some reason, I imagine a failing SSD will be easier to recover from.

Re:Industrial PC's

[NeuroManson]

True, but as far as I know, when those drives die they don't usually take your files or system with them.

Technically, however, aren't the cooling fans (basically the simplest components in the system) the most critical? If your cooling fans go out, they can take the entire system with them, and then some.

Re:Industrial PC's

[Jimmy_B]

Technically, however, aren't the cooling fans (basically the simplest components in the system) the most critical? If your cooling fans go out, they can take the entire system with them, and then some.

No, they can't, because the critical components have temperature sensors on them to protect from that. A well-built system should never toast any component regardless of fan failures.

Re:Industrial PC's

[couchslug]

Floppy drive to load RAID drivers?
Try slipstreaming instead. nlite makes it even easier.

64Gb Available Now

[Obsidian Dagger]

BUSlink 64GB flash drive part#BDP2-64G-U2 is available now with a retail price of $4,999. I checked stock before posting and there are 5 currently in stock available for drop ship from California.

The Kanguru 64GB flash drive part#KFDM-64G is priced under $3,000 but their ETA has been pushed back again to April 1.

Concerns

[StickyWidget]

Normally, I'd be ridiculously happy about this sort of development. The speed increase, the reliability increase, and the lower power consumption easily make drives like this very attractive. However, as a Computer Engineer I have to wonder if our computers and operating systems can REALLY handle a drive with a 48 MByte/sec transfer rate.

Have situations with a drive this fast ever been extensively tested in a lab to determine problems that may develop in the operating system? What happens when processes that were designed and coded with a slow hard drive in mind come in contact with ridiculously fast flash drives? Hard drive latency has been a fact of life for so long, what happens when it's no longer an issue?

Sticky "JustAskingTheQuestion" Widget

Re:Concerns

[ChronoReverse]

Raptors can easily give you 48mb/s much less really nice drives like SAS drive. And this is before factoring in RAID.

I don't how this would've been an issue in the days of DOS, much less with modern operating systems.

Re:Concerns

[burndive]

Ever hear of a RAM-based virtual disk? I use them on my Linux server for /tmp and anything else that doesn't need to persist through a reboot. They're in RAM, but the OS treats them as drives. Any Live (CD or DVD-based) distro will load the main OS onto RAM disks (if there's room) and run from there, only accessing the optical drive for large data files. It works fine.

Re:Concerns

[Anonymous Coward]

Some issues wouldn't be caught w/ ramdisks, but there have been various flash based IDE and SCSI drives for many years, so I'd say most everything would have been worked out by now in most OSs. There might be some timing issues that pop up b/c fast devices attached to IDE and/or SATA become more common and the drivers get exercised more, but I'd think that most likely you'd probably just see some stuff that performs much less that optimally using these drives and expecting them to be slow rather than stuff not working.

Re:Concerns

[suggsjc]

Yeah, but just imagine a RAID array of these...

It would still take quite a few before you would start getting close to fully saturating a PCI Express card, but just imagine if you could.

I know there will (probably) always be a performance gap between RAM and your mass storage, but hopefully this will begin to close the gap. I see many applications greatly benefiting from this type of setup/technology.

Re:Concerns

[Blackknight]

Yeah, a few of these running RAID 10 in a web server would kick ass.

The 1tb optical disk looks more interesting

[Anonymous Coward]

The article at the TFA's website about a 1TB removable disk looks even more interesting:

http://news.digitaltrends.com/article12559.html

http://www.mempile.com/

Re:The 1tb optical disk looks more interesting

[UncleTogie]

An article from MaximumPC caught my eye recently, and while it's not a TB, it IS holographic storage....

http://www.inphase-technologies.com/

Gaming

[Dancindan84]

This is definitely going to have the most application for gamers and people looking for performance PCs initially. I already know lots of gamers who go for smaller 10,000 RPM drives for their OS and applications, and have larger 7200 RPM drives for storage. This seems to be in the same ballpark in terms of size (mid-size consumer 10k RPM is 74GB), so it will just be a matter of performance vs. dollars.

Re:Gaming

[Hoi Polloi]

I wonder if this'll end up killing off Raptor drives (10K rpm). Since it is flash if the OS could save its state to the drive before shutdown we could have instant on capability too.

Re:Gaming

[maxume]

64 GB is a size downgrade I don't want, but 128 GB hard disks(2.5") are $80

http://www.newegg.com/Product/Product.aspx?Item=N8 2E16822136007

I'd be happy to pay $150-$200 for a solid state drive of the same size, just because.

Re:Gaming

[gmb61]

Assuming the drives will use the SATA interface, there should be no reason why you couldn't RAID multiple drives together to get more space. Multiple RAIDed solid state drives would have unbelievable performance.

Re:Gaming

[tsalaroth]

It's called "Hibernate" and most modern OSes support it out of the box - albeit some require some configuration to turn it on.

Re:Gaming

[NeilTheStupidHead]

And it's still not instant, since it requires writing then reading data from the hard drive. It's far faster than a clean boot, but still takes 5-10 seconds.

RAM based HD's answered these questions

[symbolset]

Years ago.

Yes, there were problems. They were fixed. All good now. Faster is better. Lower watts is better. Denser is better. Now let's work on cheaper.

The price is $496,000

[Anonymous Coward]

I came to this price from a previous slashdot article about the million $ laptop...

The laptop comes with 128GB of solid state disk space, Blu-ray, and a detachable rare diamond that acts like a power button and a security key.

High end Laptop $2500
Blueray Drive $500
Diamond $3000
128 GB solid state mem $992000 = 496,000 for 64GB = priceless!
Or is the diamond that is priceless!

http://hardware.slashdot.org/hardware/07/03/26/197 253.shtml

Performance vs 10000RPM drive?

[MontyApollo]

What would the real world performance be like compared to a 10000 RPM drive? I think their read and write speeds are comparable to WD Raptor, but I'm supposing solid state would have quicker random access. How would this compare in terms of real world performance?

Re:Performance vs 10000RPM drive?

[Anonymous Coward]

These things have a .12ms avg seek time.

Raptors are still at abt 4.5ms.

Additionally they can Read in Parallel, randomly. (ie not limited to the spindle speed and read head physical limitations.

They are NOT just as fast. No sir. They make raptors look like they're standing still, and everyone else like they're in a time warp.

I saw a demo of an XP boot / shutdown on HD vs SSD. (where'd that link go) It's pretty amazing and significant.

Add in an OS / CPU that can HANDLE parallel loading of drivers / OS modules, then it'll be something like the Linux 6 second boot (ok maybe 10 second).

Re:Performance vs 10000RPM drive?

[wile_e_wonka]

A search on youtube for 'boot' and 'ssd' resulted in this:
http://www.youtube.com/watch?v=vb1w3_KySG8

and this:
http://www.youtube.com/watch?v=SBhj56gIUuA

Perhaps that's what you're looking for.

Now listen here young man

[Anonymous Coward]

Don't get S.M.A.R.T. with me.

Young whippersnappers.

--
On Slashdot Alzheimer's means you forget who you are

Perfect solution for database?

[kamelkev]

A couple years ago we had a drive here that was essentially ram with a hard disk backup. It was only 4 gigs and was very very expensive (I can find the brand if anyone is truly curious). We got in maybe 2-3 years ago and it was like 10 grand.

We used that drive for random things, but I always envisioned putting a database on it... of course it was too small. I'm wondering if these new flash drives might be potential alternatives for us... but I'd really need to see someone talking about how many writes these guys really can do.

If anyone has *any* information that I can look at that talks about doing a high rate of writes to a volume based on these drives I would be very interested in seeing it.

please reply if so!

Re:Seek time? Should be 0ms

[26199]

Well... that doesn't necessarily mean it's as fast at random access as it is at consecutive access.

Normal computer RAM is also faster at consecutive reads than random reads.

Micro servers?

[Anonymous Coward]

Could these solid state drives be used as part of a functional micro-server farm? Cisco, for example, produces PoE switches with 15.4W on all the ports, GigE too. How much power would you need with one of the new low-power Xeon chips, 8GB or so of RAM, and one of these 32 or 64G SS drives? No HD, no spinning anything, just a serial port, and maybe a second GigE port, all in a little box. How many of these could you stuff into some chassis (just to hold them together) into a rack? 48 into 5 or 6RU?

Dude, you're on the wrong thread...

[GeorgeFitch3]

I think you meant to respond to this:

http://linux.slashdot.org/article.pl?sid=07/03/27/ 1753218

Re:Price: $500ish?

[Demon-Xanth]

Without looking around much, I see CF cards at 1GB for $12.50, 2GB ones for $16, 4GB for $34 and 8GB for $75. Making a bit less than linear progression. But you can guess about $8/GB for wholesale. 64x8=about $500.

Predict $630

[llZENll]

Hm, based on the cheapest (without rebates) memory available at $8.50/GB, figure 20% markup between the manf and retailer, thats $6.8/GB.

$435 for memory

+10% for R&D
+10% for manf (including controller, parts, etc)
-10% for manf efficency when producing 64GB/run

COST $479

RETAIL:
+20% for geewhiz-newtoy-factor/supply shortages
+10% for retail

YOUR COST: $630

sources:
http://www.pricewatch.com/flash_card_memory/secure _digital_2gb.htm

Another prediction: SSDs will offer such huge power and performance advantanges, they will sell like crazy and drop in price by a factor of 70% within 1 year from now.

Re:Predict $630

[Emetophobe]

Taken from Samsung's Webpage:

The retail price for the Q1-SSD will be KRW2.3 million (US$2,430), while the Q30-SSD will sell for KRW3.5 million (US$3,700).

Pricey stuff...

Re:Price: $500ish?

[drinkypoo]

But you can guess about $8/GB for wholesale. 64x8=about $500.

I know this is horribly silly, but if you scored some USB2 CF readers, and you've got at least two USB2 buses to dedicate to the problem, you could RAID them (stripe) and potentially get BETTER performance than with the single drive. Wouldn't help much for a laptop but it would be a super fast, super reliable storage mechanism for home use...

I know a guy who made a RAID of ZIP 100 drives on a RS6k running AIX once... Doug, where are you? I know you're around here somewhere? :)

It depends on usage models

[EmbeddedJanitor]

Sure there is no mechanical wear on flash, but the MTBF numbers for flash assume that you stay within the endurance limits.

The flash parts used in these devices can only program approx 10k times before they can be expected to start failing.

Re:It depends on usage models

[Dahamma]

The flash parts used in these devices can only program approx 10k times before they can be expected to start failing.

Modern NAND flash is in the 100k+ erase/program cycles... from an ST application note on wear leveling: "In ST NAND Flash memories each physical block can be programmed or erased reliably over 100,000 times." Of course, the wear leveling is what gets you in the 1M hour MTBF range...

the MTBF numbers for flash assume that you stay within the endurance limits.

With flash, the weak point is wear of the memory cells, in magnetic disks it's physical components like motors/actuators, heads, etc. Either way, more usage = more wear = shorter lifespan.

Besides, MTBF is not at all a good indication of the expected life of the disk - most drive manufacturers basically cheat and calculate the MTBF based on failures before any components would wear out due to usage (obviously... otherwise they'd be testing the drives for years before shipping them). So it's more of a measure of "defect rate".

An interesting comparison between SSDs and magnetic disks will be their MTBFs vs. average lifespans. I would guess with wear leveling covering the most likely point of failure, SSDs will eventually have a much higher MTBF, but also a much smaller range (I guess "deviation" would be the statistical term) in the average lifespan. It's very possible magnetic disks will have an overall longer average lifespan. But if that SSD lifespan can get into the 5-10 year range, then they are going to become REALLY popular for a lot of uses... (goodbye seek time!)

The "2nd" Moore's Law

[denobug]

As an Electrical Engineer I know that what deems to be unnecessary performance (think: 64k memory, 2GB address space, gigahertz processor) will ultimately welcomed. Even if we were to hit a performance bottlenect elsewhere we will fix them in timely fashion to take advantage of the new performance we have. This is almost as important and consistant as the Moore's Law.

Slow?

[SLi]

While 64 MB/s reads are definitely fast compared to =2.5" hard disks, can anyone explain what caps the access rate to 64 MB/s? I'd think the bus should be able to go way higher. Is flash memory inherently slow?

Re:Slow?

[ClamIAm]

Is flash memory inherently slow?

Yes.

(at least for now)

These fair similar to 2.5" drives

[Hackeron]

If comparing these to 2.5" drives instead of 1.8" drives the advantages aren't as drastic.

* 2.5" drives consume between 0.8W to 2.5W (ok, seeking eats a lot, but during sequential read or write, they consume similar amounts), almost no power consumption when they spin down.
* 2.5" drives give 53MB/sec read and write.
* 2.5" drives are very cheap and have triple the capacity.

The solid state drives are still at an advantage, but it's not quite as large as compared to 1.8" drives.

Re:These fair similar to 2.5" drives

[Astastrafal]

Access times are where flash takes the advantage. IIRC, it's ~50 ns for flash v/s ~10 ms for hard disks, thus in the order of 200,000 times faster

Here you go.

[Anonymous Coward]

http://www.superssd.com/products/tera-ramsan/

No pricing yet

[nurb432]

If you even have to ask about pricing, trust me, you cant afford it.

Re:No pricing yet

[Tom Womack]

http://www.dramx.org/#fspot gives a decent idea of pricing for the underlying chips; it's just over a dollar per gigabit at the moment, so before considering assembly charges, profit and the like you'd expect a 64G drive to cost about six hundred dollars. I'd expect them to be about a thousand dollars once all the overhead in assembling systems from chips is added in, and rather more than that initially when they've got novelty value.

Disc space hit eight dollars a gigabyte in about June 2000, and a dollar a gigabyte about last year, so give it five years ...

100k, not 10k

[Anonymous Coward]

THe Samsung site says these new drives are based on single cell level NAND technology. It doesn't have as high a density as MCL NAND. Bbut each cell can do 100K rewrites as opposed to the 10K rewrites of the more common MCL NAND. See EDN article on difference between SCL and MCL NAND http://www.edn.com/article-partner/CA6319917.html

Re:100k, not 10k

[AVee]

Well at 45MB/s write speed it would take about 4.5 years to rewrite every bit 100k times. So as long as the usage of the disk doesn't have a nasty hotspot you'd be fine. But on the other hand, you should be able to 'burn' a targeted 1MB section of the drive within an hour. There are sufficient usage scenarios where a certain section of a disk is updated regularly. Updating the same bit every second will give it a lifespan of about a month, that is seriously short.

Not the MTBF, the read-write cycles.

[DrYak]

It's not about the MTBF (the wear with age), yes you can almost indefinitely read data from you flash drive, when compared to harddrives, because there's no mechanical wear.

BUT!

The flash cells have a limited number of write cycles, which is very small compared to hard drives. If you write too much data on the same sector, the sector get very quickly broken.
If you used a flash card for swap, it won't last long at all (because some sectors get constantly written over).

To limit those damages, flash controllers use "wear level". That means that the small RISC controller that interface between the flash cells and the computer interface (ATA/CF, SD, USB, etc.) dynamically remaps the sectors so the wear caused by write cycles is distributed over several different sector.
Let's say that an OS constatly writes data on the first couple of sectors. Instead of always writing on the first few cell, the controller remaps a different physical flash cell, to the logical disc sector seen by the OS.
This works as a charm for flash media storing files likes used in digital cameras and such.

But doesn't perform as well when used by an operating system.
Windows XP is specially bad at this.
Other OS - such as Linux or *BSD, that already have good support for running on slow read-only media (LiveCDs) for a long time, that don't need writing that much (except /var and /tmp, most of the rest of the installation can be read-only), and that support special file systems designed for lower wear (JFFS and such), may fare better : for example there are some Linux distribution that are tested for running from flash, like Damn Small Linux.

Re:Not the MTBF, the read-write cycles.

[Anonymous Coward]

Something else worth noting, once the write-cycle limit is exceeded and the sector is bad on the flash device, it is completely unrecoverable. Whereas when a magnetic disc sector goes bad there remains a decent chance that data can be recovered.

Re:Not the MTBF, the read-write cycles.

[Deliveranc3]

But doesn't perform as well when used by an operating system.

Since flash doesn't have sectors that are faster than others; Thus, this is incorrect.

Flash chips each have a read and write speed limit the more of them you have in parrelel the faster it can read/write. It's trivial to make the chips within a flash drive have JBOD(F) properties.

This is one of the major advantages for me, disks that will be able to max out gigabit+ ethernet with increadible seek times, data redundancy, and massive througput.

As disks get bigger it may become nescessary to have some space for a read/write buffer (normal HD's have ram for this) which will increase the life or need for higher MTBF sections, both of these properties are showing up in variations on flash.

So if you have a flash disk with 1 Increadible MTBF chip, 1 super speed no storage sector (like ram), 1 massive storage space, and a bunch of standard flash you can have all the advantages of every kind of disk with the internal controller handling performance and wear leveling (not a trivial programming problem but one which we have a bunch of excellent solutions in place for).

My personal problem with flash disks is that industry seems to be holding back development, trying to develop an upgrade cycle instead of realeasing a perfect solution.

I can get a 1GB microsd flash card for $15 about 400-600(conservative) of them would fit into a 2.5 disk enclosure. With JBOD and wear leveling across the chips and I'm assuming it would be cheaper because you wouldn't need hundreds of cases/interfaces a 200GB drive with read/write speeds of 100-300 Gb a sec and seek time of
Hmm, well maybe the price does need to come down but the other concerns about flash seem unjustified, write wear isn't a problem, it's not scary. losing all your data to a HD failure, now that's scary.

Re:Not the MTBF, the read-write cycles.

[frdmfghtr]

Other OS - such as Linux or *BSD, that already have good support for running on slow read-only media (LiveCDs) for a long time, that don't need writing that much (except /var and /tmp, most of the rest of the installation can be read-only), and that support special file systems designed for lower wear (JFFS and such), may fare better : for example there are some Linux distribution that are tested for running from flash, like Damn Small Linux.

At the risk of asking a dumb question or missing an important point...Would you mitigate this wear by creating a RAM disk for items that are frequently read/written, and periodically (every few hours? When going to standby? Only on shutdown?) update the flash drive with the "temporary" contents. I can see the argument that a swap file in RAM is pointless (simulating virtual memory on a RAM disk would be silly--if you have the RAM for a swap file on a RAM disk, you have RAM to use as RAM directly) but things that get frequently written would be RAM disk candidates (home directories for example?)

Just a thought.

Re:Not the MTBF, the read-write cycles.

[Anonymous Coward]

The flash cells have a limited number of write cycles, Yes, but... how many of you are going to keep a computer system for more than a few years? Not many I imagine. With CPU/DDRAM performance improving every 12-18 months, a system has a life expectancy of less than three years. So, if the flash RAM lasts only 5 years it ain't no big deal! Now, if I can load a COD:UO map in 5 seconds instead of 30, whooho! Sign me up!

Re:Not the MTBF, the read-write cycles.

[fuliginous]

I suppose an answer to this for a Linux system in regular desktop use is to slap in 2, 4 or more Gb of regular memory so that you don't have to use swap and can perhaps mount 1G or so as the regularly written drives.

Well

[ShooterNeo]

Obviously, from a technical standpoint, flash drives should ALWAYS be faster than mechanical ones. MUCH, MUCH faster. That seek time, which is a fraction of the time that a hard disk takes, shows that the electronics can get to their data quicker. The catch is the electronics in convential flash disks have been designed to drive very small drives, and so there are bottlenecks that can make their transfer rates slower. However, in theory flash memory can be read in parallel and have a transfer rate of "the sky is the limit".

Simple solution!

[Anonymous Coward]

Put your swap partition in a ram disk!

Re:10 reformats/day for 27 years!

[Anonymous Coward]

The Samsung site says they are using single cell level NAND. This has a 10 times the rewrite capability of multi level cell NAND found in other flash devices. That's about equivalent to reformating the entire drive 10 times a day for 27 years. Reliable enough?

Re:Price: $500ish?

[tsalaroth]

I'm sorry, Doug is in the server room, swapping Zip disks. How can I help you?

SATA drives

[Anonymous Coward]

I'm still waiting for an internal SATA connected flash card reader. Just plug several SD cards into it and go. This would provide great flexibility and you could buy flash from the cheapest company instead of going to the one or two that make these things.

Seems like a perfect way ...

[constantnormal]

... to cool down those warm AppleTV devices.

Now if Intel will just ramp up their phase-change memory alternative to flash, and get it out in a comparable size, the issues regarding cycling limits will be dealt with.

Wont be an issue

[spagetti_code]

There are two reasons why this, in theory, could be a problem, but in reality wont be.

a) Guaranteed writes for flash cells are now in the millions
b) almost all (and possibly all) flash memory systems use
write levelling technology to ensure the write load is spread.

We use them for small 24x7 computers doing UI and data capture
work, and after several years the flash has yet to fail on any of them.

Compressed pagefile

[tepples]

If your pagefile exists in RAM, then why even have a pagefile at all? Just get rid of the pagefile completely. In the mid to late 1990s, Connectix sold RAM Doubler, a virtual memory manager for Mac OS (classic) that used a compressed pagefile. The idea of the product was that LZ77 compression and decompression are faster than seeking.

mnb Re:Perfect for MP3 players

[Anonymous Coward]

The hard drive consumes more power than the rest of the DAP (sans backlight) when it is running.
There is a reason most modern DAPs have 32 or 64MB of DRAM for buffering.

FWIF, my iPod Mini has seen no appreciable battery life improvement since switching from the microdrive to a CF card. A controled experiment might show otherwise, but I haven't noticed anything while using it normally.

Re:mnb Re:Perfect for MP3 players

[Grishnakh]

You should check on the power consumption numbers for the microdrive and your CF card and see how they compare. Remember, the microdrive is quite a bit smaller than the 1.8" drives used in most larger players; it probably takes a lot more power to spin up all that extra mass.

You could be right though; the only way to really determine how much this would help play time is to try it out I think.

Ok, but not measured in milliseconds.

[CFD339]

Sure, there are differences but we're talking about nanoseconds, not milliseconds with respect to ram. At worst it will be orders of magnitude better than a spindle and head.

Re:Ok, but not measured in milliseconds.

[26199]

Hmm, yes. I've no idea why that coment was worthy of a +5. Ah well.

A few replies a bit further down seem to have actually answered the question, maybe they'll get modded up in due course...

Just put in twice as much RAM....

[itsdapead]

... and drastically reduce the load on the pagefile.

Seriously, the way of the future is surely to loose these distinctions between RAM and mass storage. Just leave it up to the OS to juggle data between CPU Cache, fast RAM, slow RAM, Flash and the internet, and refer to every bit of data by its IPv6 address... :-)

Big Time Data Stream

[Anonymous Coward]

I need something to suck up 1.5 G/s which is what the HD SDI standard is. Those uncompressed HD 1080 30p streams are insane. Any ideas?

  I have an account but they broke it and can't fix it ... heh.

When it fails

[Colin Smith]

You recover your data from your last backup. Just as you do when your bog standard sata drive fails.

Re:When it fails

[HTH NE1]

And if your backups are bad, or you lost data entered between the times you perform backups?

The question is not what methods do you have to restore data and how you can protect yourself; it is what expertise exists out there that can restore data from worst case scenarios involving flash memory like is available for magnetic media.

And also what business opportunities exist for those seeking to provide such service, because there will be people who will need such services.

Latency ?

[Yvanhoe]

Surely the latency is the best advantage of solid state vs hard drives, why isn't it presented in the summary ? It must be impressive like a few usec compared to 200 ms

Two Thoughts

[wintermute1974]

(1.) Where do I put my swap file? If I use flash for my swap, I imagine my MTBF on any flash drive would be measured in months, if not weeks, for PCs that are never turned off.

(2.) How many motherboards will die sad deaths due to these ultra-fast drives using HD interfaces? I'm sure a lot of chipsets are only tested on hard drives, not on the technical limitations of the bus. Failures never seen in development will happen in real life.

A better use?

[Plekto]

I can see these easily replacing all of those Raptors and being huge in a typical server environment.

Consider that with a raid array, failures are rarely catastrophic in nature. But the time wouldn't be months (mostly due to heat killing the drives quicker), but years. With these, though, you'd save enormous amount of power, have much less heat, require smaller power supplies, and so on.

Imagine a raid array at a major company. 200 drives. 1KW power supply driving it all.

Re: Performance

[migmog-gomgim]

One problem is that filesystems are designed for the physical characteristics of a disk. If you want to overwrite a sector on a disk, you just point the head at the right place and write.

You can't do this on flash.

You can write minimum a page at a time. With parallelism in place this may be something like 32KB which is similar to cluster size on some filesystems so no problem there. However, you can't OVERwrite existing data. You have to write it somewhere else, and have control structures in place to know where it was put, and that the old data is now invalid. Then you have to consolidate the data in order to get the now obsolete storage back. You can only erase a block at a time, which could be in chunks of 1MB or so with parallelism.

So to change some text in the middle of a document, or swap a page of RAM to virtual memory, you will actually be doing CONSIDERABLY more actual reading/writing/erasing than the few bytes that the OS/filesystem assumes you're doing... because it's designed to use a disk.

So the performance numbers you see are for sustained sequential reads and writes. I'd be more interested in 'real' performance numbers for PC based operation. And... while this is a common caveat for disk benchmarks, it is WAY more of a significant issue for flash disks.

BTW, for just a few dollars you can make your own (small) SSD by purchasing a IDE-CF converter and plugging in a compact flash card. It's fast, cheap, silent and low powered. Ideal for your OS/boot partition. Just don't put a swapfile on it!

MacBook mini or ...

[MrHatken]

would that be nano?

Look out for these to be in the upcoming MacBook mini/nano (carbon fibre black)!

If only I could get my apps and files down to 64GB :-( I'm moving my laptop up to 160GB, it's easier than walking through all the files to make space.

Cheers,
Ashley.

Link to actual Samsung news item

[xocp]

FWIW, here is a link to the actual Samsung news item.

From the article:

Samsung plans to start mass production of the 1.8"-type 64GB flash-SSD in the second quarter of this year.

One Upping.

[camperdave]

Well, when I look at my list of the powers of 2 (which is sorted in ascending order), 64 is one up from 128...

That is going to rock when the price comes down

[Anonymous Coward]

This is very cool technology. When the price comes down, man, we'd be crazy not to go for it. I'd love to have a laptop with this kind of "hard disk" in it; it's way more shockproof than a Winchester hard disk.

Technology is so cool.

mailto:swalker@cmosnetworks.com

How reliable does it have to be?

[Anonymous Coward]

If they became widely used and produced as a system drive, how easy is it to ghost something that's only 30-60gb onto a larger secondary drive (IDE, SATA etc)?

Even if it only lasted a year or so, if they were inexpensive enough you could pretty quickly drop a new one in and re-image it from your backup in very quick time.

Correction

[Emetophobe]

Er, I'm an idiot. Those are the prices for a different Samsung product with solid state drives included and not the price of solid state drives by themselves...

You're just plain wrong here.

[CFD339]

Admittedly not about noise -- notebook drives aren't too loud -- but they are slow, and they do use a lot of power. A 25% power increase and a very big speed gain would work wonders for me. Yep, I'd pay it.

My favorite line

[fishbowl]

IANAL, but I have studied law and I have worked in the litigation field. I have read many letters that have had me wanting to ROFL, and this is in that category. But the best part is also the last bit:

  "From there, it should be a short trip to dismissal even if it means getting our clients to mediate Mr. Merchant's positive claims in the absence of an appropriate settlement."

Translation: If you have read this far, you realize that you not only have no case, but that you are entirely out of your league because the standards of evidence in the court system where I have major influence, would procedurally bar you from even entering your case on the docket. Despite this, my client's claims against you are already demonstrated, and our claims will continue to have merit even after your case is dismissed with prejudice (and we have not offered to drop our case.)

This letter is a masterpiece because it manages to hand the plaintiff his ass, in a rather respectful colleague-to-colleague way, while at the same time threatening a counterclaim that could end up with far greater damages than the initial claim!

And the real beauty is that even though the RIAA seems to have withdrawn its claim, the damages from the malice might still hold, if they really want to push it.

Who did they sue? Directors of a Silicon Valley bank? They should do some research before they pull the pin on the hand grenade!

"I would be happy to send the airplane..." (At the plaintiff's expense of course...)

Love it.

RAM: indeed

[DrYak]

At the risk of asking a dumb question or missing an important point...Would you mitigate this wear by creating a RAM disk for items that are frequently read/written, and periodically (every few hours? When going to standby? Only on shutdown?) update the flash drive with the "temporary" contents.

Yes indeed, that's part of all optimisations done in LiveCD. Specially using "union" type of mounting where several filesystems are mounted on the same point (when read, data is pulled from the CD-R, but then, subsequent modification go to RAM disk) and similar solutions.
Also, as I said, the often over-written zone are limited (Linux doesn't write much on disk when it isn't needed) and this makes easier to use such solutions.

I can see the argument that a swap file in RAM is pointless (simulating virtual memory on a RAM disk would be silly--if you have the RAM for a swap file on a RAM disk, you have RAM to use as RAM directly

RAM used as RAM : and the system could use it even more efficiently.

But swap on RAM disk isn't completly silly... if it's a *hardware* RAM disk :
once you've maxed out all memory slot on your motherboard (say, 2GB DDR-2 dimms in each slot), the only solution to keep adding more memory is to of those "conver RAM dimms into a SATA harddisk" solutions like the Gigabyte's iRAM that was featured on /.
It won't be as fast and directly usable as the main DDR-2 memory, but it enbales you to add more memory to the system and, if you put your swap file on it, in the end it does extend the maximal memory limit, although in this cas it's *virtual* memory and over a slower connection.
It's kind of "double the number of memory slots, although newer are slower"

Run applications directly from flash ?

[Coati]

In embedded systems with small memory footprint we usually run code directly from flash.
What about doing it in personal computers as well (without the ATA interface of course) ?
This would free a lot of RAM and we still have the CPU cache to speed up the access.
What would be the required access time / transfer rate to start thinking of it ?

I am convinced that we don't need all that ram (640k is more than we will ever need ;-) ). Even all the constant data could be read directly from flash (we do that in embedded systems).

Of course. It would require changes in the hardware architecture and in the software programming. A change, which would be easier to do with an open source operating system...

Second drive

[Moraelin]

So use it as a second drive.

I, for example, would cheerfully buy one of these for my gaming desktop as drive D:. Well, ok, E:. My WD Raptor is plenty fast to boot Windows. I have no real need to accelerate that, so C: can jolly well stay a hard drive.

Load times of other stuf, however, can suck plenty even on that.

E.g., games. Games are a prime candidate for this kind of thing, since basically they don't change often. A game might get pathched a few times, maybe even once every couple of weeks for a MMO, but otherwise tends to be left alone. Most nowadays also save their save games in some subdirectory of "My Files", not in their own directory. So the game drive wouldn't even be worn out by my quicksave addiction.

HTTP/1.1 302 Found

[Looce]

Location: http://yro.slashdot.org/article.pl?sid=07/03/28/01 11205

You may want to put your comment there. :)

Re:HTTP/1.1 302 Found

[fishbowl]

Oh crap, thanks. Million years on slashdot and I have made some boneheaded comments, but never before plain posted to the wrong subject.

The swap/pagefile is NEVER used

[Anonymous Coward]

Okay, what is it about these swap partitions that make you people think they are actually used?

System 1: 1GB RAM (13MB free), 1GB swap (0 used)
System 2: 1GB RAM (140MB free), 1.5GB swap (0 used)
System 3: 128MB RAM (17MB free), 230MB swap (0 used)

Swap is STRICTLY for when you run out of memory due to a memory leak, of if you are doing some reallly memory intensive processing -- the kind of processing that takes years.

You guys must be using broken operating systems.

Re:10 reformats/day for 27 years!

[Silverstrike]

No?

Assuming you're talking about doing a complete format, ie: zeroing all the bits, and not just whacking the file directory, then a "format" still only implies that you're writing all the sectors once per format.

27 years * 365 days * 10 = 98550 writes = ~100,000 writes

That's not a lot. Consider if there's a temp directory that is being written to 1/second. There's 86,000 seconds in a day. That sector will be dead in just over 1 day.

In the end though, "10 formats a day for 27 years" is a meaningless statistic. As several people have pointed out, there are controllers switching the frequently written sectors around. I didn't read the Samsung site, nor do I care to. It just struck me as a geek version of the "5000 song MP3 player" marketing lingo we hear all too often.

Skip Swap...?

[pentalive]

If your going to design a machine with this kind of drive in mind wouldn't you design it to have a large main memory, so the need for swap is reduced?

While your at it, if this is to be a server class machine and money were no object (whups that's a BHIG if) the I would have a large DRAM cache between my flashram HD and the system to cut down on the number of read/write cycles.

OOh yeah!

[diorcc]

Can't wait to get my hands on this...
Definately OS + apps on it only (mostly),
the rest of the files (videos etc) can go on regular drives.

bigger picture

[COMON$]

easy backups, imagine being able to restore servers in a single bound!

Vista ....

[tinkerghost]

Vista is using a 'smart cache' that is supposed to link to flash thumbdrives - in theory it sets the swap file up like L1 & L2 processor caches, with the L1 cache going to the thumb drive & the L2 to the HD. I am assuming that if they've implimented it as a standard for a 'business class' operating system, they've at least done some feasability studies here - though I know for a fact that most places with a security clearance requirement won't be able to make use of this feature so I could be wrong about that assumption.

Re:Seek time? Should be 0ms

[apocalysque]

granted, there is (vitrually) no seek time with SSD, but most people think that seek time is synonomous with access time and comparing the access time of a SSD with the seek time of a HDD would be a good way to compare the two.