Solid State Memory on the Rise

skaet writes "CNet is reporting that manufacturers of NAND flash memory are expanding the market for their chips - over the next few years - to eventually replace current methods of storage in media capture devices, mobile phones and even some notebooks as well as car navigation systems and large data storage at corporations and government agencies. From the article: 'The average notebook has 30GB (of hard drive storage). How long is it before the notebook has solid state memory? Five or six years,' according to Steve Appleton, CEO of Micron Technology, one of the world's largest memory makers. 'I'm not saying drives will go away. There will always be a need for storage, but when was the last time you tapped out a drive?'"

Comments

Filled up a drive?

[DietCoke]

This guy clearly hasn't ever installed Bittorrent.

Re:Filled up a drive?

[ScrewMaster]

Yeah, no shit. People talk about how Bit Torrent downloads are consuming some large percentage of the Internet capacity ... but it's also consuming a larger portion of user's hard drives. All RIAA/MPAA bitching aside, downloading is driving the sale of a lot of storage hardware. It's even worse if (like me) you're a packrat and just hate to throw anything away. So, you don't ... you just buy a couple more hard disks and jack them into your RAID array.

Re:Filled up a drive?

[Wyatt Earp]

Hell, I don't even need to use Bittorrent to fill drives.

Grab some DV streams, then edit some iMovie and save some complete iDVD projects and it fills up fast. It'll get even more size intensive when HD capture devices come more prevalent.

Re:Filled up a drive?

[GuyverDH]

LOL - Try converting your movie collection to digital storage, for a media center system.

3.5 TiB and climbing.

Re:Filled up a drive?

[DietCoke]

Agreed. I copy *my* DVD's directly to my HTPC, vobs and all, because I'm too lazy to search through a rack of DVDs to decide what to watch. Sure, I could rip em and repackage 'em using Xvid, but the point is convenience.

Do I *need* more than 30 gigs of space to live on? Well, no. But life sure is more entertaining and easier when you're tapping out at 1 terabyte, rather than 30 gigs.

Another reason this idea won't work, either: imagine the environmental costs. Making chips is dirtier in terms of byproducts and materials, whereas hard disks are relatively easy to break down. I can't see this helping companies trying to attain ISO certification.

Re:Filled up a drive?

[DietCoke]

You read my mind :) Just posted that as a followup.

3.5 gigs? Sheesh. My wife would kick my ass.

Re:Filled up a drive?

[DietCoke]

Now imagine that 250 TB was all solid-state. Even with more memory going on those chips now, still... I'd be surprise to see it match a hard drive in terms of physical size, bit for bit.

A hard drive from 1986, maybe.

Re:Filled up a drive?

[Wisgary]

I sometimes look at my download folder in awe, it's full of so much useless SHIT that I have no use for (or ever will have a use for, since there are new versions of just about ANYTHING in there) but... sometimes... it's hard to hit that delete key. It really is, I think we have a new symptom of obsessive compulsive disorder. I wonder how long until psychologists start to ask... "How long has it been since you deleted stuff from your download folder?"

Re:Filled up a drive?

[empaler]

I have the same problem but found a fix for it a short while ago - my laptop had to go in for service and I was too lazy to burn more than a single backup CD.
Also, I don't think my downloads directory is any business of the service technicians (and, as we all know, they do look at your stuff, especially if they're bored) - so I wiped the entire Documents folder and generally scrubbed my computer for personal data.
I saw it as a healthy practise, both from the standpoint of my private data but also because I had so much shit.

Re:Filled up a drive?

[Schraegstrichpunkt]

It's even worse if (like me) you're a packrat and just hate to throw anything away.

Sigh. Really: You don't need *all* of those hamm CD images on your hard drive.

At least delete the m68k binaries you never used anyway. If they're so important to you, burn them to DVD or something. ;)

Re:Filled up a drive?

[LiquidCoooled]

I just upgraded my computer, and have left behind a 200gb drive with my life.
Its got EVERYTHING small I downloaded and did for the last few years (last time I formatted and installed must have been 2001).
All my emails and all my favorites and all my documents - GONE (well, put into offline store)

The last time I did anything like this was when my Amiga died - the harddrive is still around somewhere though.

I go into the old stuff so infrequently that I doubt I will need to dive into it often.

I did however bring the other 200gb drive with the media shares, and I thought about which applications I need, but the historical downloads were just bloat.

Re:Filled up a drive?

[Phantom Zmoove]

I suffer from a similar problem except it bleeds out from data into life. I hate to delete real things. (or even sell them) It drives my wife crazy. So my solution is to just to be very picky about anything that I ever bring into the house. It takes me so long to make a decision to buy something, because I know that I will have it for ever.

Its not a permanent solution to the problem, but it has slowed its progression down a bit.

Re:Filled up a drive?

[nacturation]

Wow, Titanium Bytes. Now that's serious storage. Oh, wait...

Re:Filled up a drive?

[jacksonj04]

I have two drives (a 20gb and an 80gb) with 'old' stuff on, and I just keep them in a network storage array. I can then use Google Desktop (Or similar, ymmv) to quickly run through what is there if I need something. The advantage is I can keep the file structure of the old systems exactly the same, and find (for example) fonts I use when designing. The disadvantage... I've got 100gb of crap sat on my network :(

Re:Filled up a drive?

[InvalidError]

Hehehe, I know that tune.

Every time I buy a new HDD, I tell myself I will screen all the stuff I move over to it and use the remaining space to re-organize my other HDDs... but always end up filling up the new drive with new stuff and having to buy yet another new drive because I no longer have enough spare capacity to shuffle my older drives' contents. There must be well over 100GB of stuff I could safely delete but simply do not care to when a new 300GB drive costs only a byte over $100.

Next time I need storage, I think I will look into an eSATA RAID5/6 box with 5/6x300GB... I'll be damned if I manage to fill that and still not manage to do my big re-org before running out of space again - moving the ~1TB I have scattered across PCs and external boxes on the array would be much less painful than connecting a few of my external drives to assemble collections to put back on another drive, repeating for each drive where I want to change the file collection, usually to put stuff I am most likely to need on my newest/fastest drives and dumping stuff I may never use again on loose older drives I dump in anti-static bags between uses.

Re:Filled up a drive?

[InvalidError]

I hate that too.

Old hardware has little to no resell value and everytime I am told that I should get rid of my old PCs, I tell myself and whoever asks:
1- This old junk is not worth selling
2- Whenever I need old hardware for experiments, I would be asked to pay a small fortune (compared to the price of not dumping/selling my old stuff) to get some, I am better off keeping spare oldies in case I kill one
3- I still have some vacant storage spots for old stuff

As far as computer stuff is concerned, I also try to put it off until absolutely necessary... for core PC components, this usually translates into better, more mature and cheaper parts. I usually try to stretch my main PCs to 3-4 years, that helps to keep the number of 'dark boxes' increase slowly.

As for TFA's 30GB average laptop, seems like they have not looked at the retail market recently - 80GB is mainstream right now. That 30GB average makes it read like their target market is HDD replacement for ~3 years old laptops.

Re:Filled up a drive?

[osu-neko]

As for TFA's 30GB average laptop, seems like they have not looked at the retail market recently - 80GB is mainstream right now.

Well, they said the average laptop, so I assume they mean the average laptop. Looking at the retail market only tells you what the average laptop purchased today has, but one assumes the average laptop is more than one day old. Mine is over a year old, for example. The average laptop is certainly less impressive than the average laptop available for retail purchase.

Re:Filled up a drive?

[GuyverDH]

Tebibytes - the proper term for measuring binary storage, as opposed to terabytes.

Re:Filled up a drive?

[GuyverDH]

That's one reason why I married a shorter woman.

Re:Filled up a drive?

[MechaStreisand]

It's not the "proper" term, sir. A lot of people - most people, probably - still use the old terabytes, gigabytes, megabytes, and kilobytes to mean 2^40, 2^30, 2^20, and 2^10 bytes, and we don't like being told that we're wrong for doing so. Some people even think that the hard drive companies were wrong for first abusing the meanings of the terms. The so-called SI prefixes never meant the same thing in computer science that they did elsewhere.

Plus, "tebibytes" sounds rather ridiculous.

Re:Filled up a drive?

[GuyverDH]

Just because everyone does (or a lot of people do) something, doesn't make it right.

The metric system is all base 10.
The binary system is all base 2.

The only 'proper' way to measure binary, is with a binary based system, not a decimal based system.

By using these terms exclusively, how many people would be hoodwinked by the storage industry?

Hell, how many people get their brand new 512MB or 1024MB flash memory devices home, only to find out that they are really 487MiB, or 975MiB in size.

Whether or not you like to be told you're wrong by using the misleading terminology matters not. It's a statement of fact, that it's incorrect terminology, regardless of historical usage, or in this case, mis-usage.

As far as sounding ridiculous, how would you like to use the long hand version of terabinarybytes? I know I'd sure hate to type that out, so tebibytes works just fine for me.

Re:Filled up a drive?

[MechaStreisand]

No, sir, it is NOT incorrect to use MB to mean 2^20 bytes. Just because SI uses Mega, Giga, and etc to mean powers of 10 does not mean that it is incorrect for computer science to use them to mean powers of 2. Computer Science units are not SI, and so just because they don't fit in with SI prefixes, that does not mean that they are incorrect. You are in the wrong for thinking that they must abide by SI rules.

Re:Filled up a drive?

[GuyverDH]

I tell you what.

Why don't you just go ahead and use whatever you'd like, and never know for sure just *what* you mean, while I use the term that means exactly what it's supposed to.

I will be able to carry my meaning through in conversation and documentation, while others will scratch their heads and wonder just what exactly you meant.

Please recall that using terms like megabyte, gigabyte when discussing storage does NOT mean 2^xx - it means exactly what it says, 10^xx.

A 500GB drive does NOT contain 500GiB or 500x1024x1024x1024, it contains 500x1000x1000x1000 bytes. Say what you want, but the numbers prove out.

And even then, that's not exactly true.
For in the 3.5" floppy medium, the 1.44MB meant 1.44 * 1024 * 1000 - so where's the consistancy?

Re:Filled up a drive?

[InvalidError]

Flash HDDs are not going to become mainstream until they match the GB/$ figures and general storage capacities. 30GB is enough to build a usable box but few people would buy a 30GB Flash drive when there are much less expensive 120GB mechanical drives around - unless they absolutely require the low-power or ruggedness.

Slow

[Jtoxification]

Most solid-state memory is pretty darn slow, and the stuff that's fast costs major $$$ ... I'll buy it when it gets faster & cheaper - but then, flash *is* much faster than the ol' floppy - I was glad to see that go ...

Re:Slow

[djupedal]

Most solid-state memory is pretty darn slow...

I was once asked to demo a solid-state HD...built with nothing but DRAM. This was a decade ago, and it was only proof-of-concept. It was only 2gb, but it would format instantly. Don't confuse SD and CF cards with DRAM. Micron makes DRAM.

Re:Slow

[Rosyna]

It was only 2gb, but it would format instantly.

Well, uhm. Formatting is almost a no-op. It just deletes a very, very small set of information about the volume and locations of files. They don't often delete any of the actual files (unless you do a low level or zero all data). Hell, even many floppy disks format "instantly". A real test of speed is read/write speeds, not a simple format.

Re:Slow

[Ethan Allison]

Well 2gb a decade ago was nice for a traditional hard drive. :P

Re:Slow

[core plexus]

If that is slow, then maybe what you want is gold nanorods. This could be the next generation in optical chips.

Re:Slow

[boingyzain]

You can't be serious. Solid-state memory is MUCH faster than normal hard drives. The read/write speeds of solid-state memory are close to that of RAM due to the fact that there is no hard drive "head" that needs to physically move around the platter.

Re:Slow

[StarWreck]

Assuming that you're referring to Flash Memory. Its getting faster, fast.

Flash memory is currently using the same speed ratings as a CD-ROM does. 1X == 150 Kilobytes per second

Secure Digital Flash memory is commonly available in speeds up to 150x. 22,500 Kilobytes per second.

We're already starting to see 200x: 30,000 Kilobytes per second.

I can boot an operating system, Knoppix Linux, with a full graphical user interface, full hardware support, multi-media, and office applications on an old 24x CD-ROM without "too much" discomfort, I imagine booting it off a 200x flash card would be relatively comfortable.

Re:Slow

[djupedal]

A real test of speed is read/write speeds, not a simple format. I didn't say 'simple format, you did :) And my example was just one out of several where the solid-state HD routinely outperformed platter based mechanisms.

Ever notice how your system seems to slow to a crawl when pageouts kick in, and the OS starts using the HD instead of RAM? DRAM is always faster than platters. Again, flash memory is slow. DRAM, which isn't.

What type of buffer (8 & 16mb) is used in those 'fast' new hd's? Solid-state, of course. Not because it's slower, that much is certain. Want a full format? Pull the power...takes only a few seconds for the data to bleed down and a clean slate to be ready for use.

Re:Slow

[djupedal]

As mentioned, tho, cost is an issue. Just three weeks ago, I went out shopping for a 1GB stick of RAM and a 250GB/7200rpm/8mb cache HD. I paid exactly the same for each item... $106.50

Re:Slow

[penguinboy]

What type of buffer (8 & 16mb) is used in those 'fast' new hd's? Solid-state, of course.

No kidding? I thought they used vacuum tubes.

Re:Slow

[adrianmonk]

Most solid-state memory is pretty darn slow, and the stuff that's fast costs major $$$ ... I'll buy it when it gets faster & cheaper

As a guy who works on apps for Palm OS for a living, I've learned that flash memory has two really nice properties that hard drives don't have:

1. Its access time is pretty much negligible. There is no head that has to be moved across the disk. Sure, there are bound to be advantages to one large read (or write) compared to several smaller ones, but the penalty for reading from (or writing to) different spots all over the place is way, way smaller than it is on a hard drive.
2. Probably more importantly, flash devices can come out of power saving mode much faster than hard drives can. This is for one simple reason: when a hard drive goes into power saving mode, it has to make a big change in angular momentum of the platter in order to come out of power saving mode. Since the penalty is so high, you have to make a compromise: either you must use more energy and keep the drive powered on longer, or you must wait for sometimes 5 or 10 seconds just to get a single byte off the disk. With flash, you don't have this problem, because it takes more like 1/2 second or less to bring the thing out of power saving mode to full functionality.

#2 is such a big benefit that I'd really like to have a laptop with a few GB of flash memory that acts as a read and write cache for the hard drive. With a good caching algorithm, it should be possible to keep the hard drive spun down most of the time and save a ton of energy.

Re:Slow

[Anonymous Coward]

Search for "ramsan" :

The RamSan-400 hooks up via a 4GB Fibre Channel interface and has 3000 MB/s random sustained external throughput. That's three gigabytes per second. Sustained. Importantly, it can also handle 400,000 I/Os per second, with a latency of less than 15s.

Re:Slow

[pyrotic]

I've seen combined disk/RAM drives in servers - at work we've got a 140G RAID 1 array with 192M battery-backed write cache. Write performance is so good we've stopped bothering worrying about filesystem optimisation.

That's a great idea for laptops, as you have battery built in, and spinning down disks saves bettery life. So you'd have 2G RAM, 4G slower solid state disk cache on the ATA bus, then 100G hard disk on the same bus with a bit of software to deal with it. Just hope you can fit enough usefull stuff in 4G, what with modern software bloat.

Re:Slow

[MrLizardo]

actually booting off of even a "slow" flash memory device, like an older usb drive will be quite quick. Much faster than booting off of a CD-ROM and quite close to the speed of booting off of a hard drive. During a normal boot process you're loading a lot of smallish programs/files, and this plays to the advantages of flash media: no seek times. CD-ROMs have seek times in the tens of milliseconds (maybe even 100 ms for an older unit). Harddrives less than 10 ms these days. Flash media on the other hand is truly random access in the same way that DRAM is, in that there isn't any kind of "seeking" done.

Re:Slow

[Jherek Carnelian]

I'd really like to have a laptop with a few GB of flash memory that acts as a read and write cache for the hard drive. With a good caching algorithm, it should be possible to keep the hard drive spun down most of the time and save a ton of energy

You are not the only one thinking of that.

Re:Slow

[AtrN]

Formatting, for most file system structures, is dependent upon seek times and with solid-state storage seeks are non-existent. Hence the speed up.

Re:Slow

[DrSkwid]

cost is not the same as price

Re:Slow

[ebob9]

Dosen't Flash memory have a limited amount of writes?

I worked for a company that made a prototype network appliance, and they used flash memory for the filesystem. 3 months after beta started, all the devices failed due to the swap-space on the Flash chips writing too often, and causing the flash storage to fail.

Re:Slow

[Andy Dodd]

To avoid the nitpickers next time, you might want to say "solid-state nonvolatile memory is slow", since even though I'm sure you meant nonvolatile memory such as flash, everyone thought you were lumping in volatile memory (DRAM, SRAM).

It's too bad drives based on volatile memory died out. There are always 1-2 models available, but they invariably use the latest and greatest memory, which is expensive.

It would be nice to be able to mix memory types in a system - Have 512M-1 gig of the highest quality superfast stuff, but then also have even more older and slower memory. (In essence, adding another level of caching). A system with 256M of PC3200 and 4 gigs of memory running at a speed equal to PC100 would most likely blow away a system with 1 gig of PC3200 in memory-intensive applications. (Yes, I know that right now, older memory is more expensive per megabyte than newer stuff due to economies of scale, but if the demand for older memory were much greater, there would be production of the older stuff at rock-bottom prices.) If flash memory can reach prices below $50/gig, older DRAM technologies should be able to achieve half that in volume.

As to flash memory - It's going to replace hard drives in some niches (it already has) where the storage required will likely not go up much from current requirements. (Automotive navigation systems for example - there's not much data you can add beyond what is already required. Compared to a good street database, POI databases are tiny and that's about all they can add.) It's never going to replace hard drives in most places HDs are used. Just like everything else in computing, new applications arise to take advantage of advances in technology (Examples - media servers where someone's entire DVD collection is stored on hard drives rather than DVDs for convenience of access.) See one of the first poster's comments about this guy clearly having no knowledge of BitTorrent. :)

Also, like you said, flash memory is slow. To those who mention the existence of 150X and higher SD media - Keep in mind those are absolute maximum ratings. Read speeds can easily go that high, but write speeds of almost all nonvolatile solid-state memory technologies are waaaay behind their read speeds (except for flash cards with very low read speeds - read speeds keep improving but write speeds haven't improved that much), or even the read/write speeds of hard drives. There is research being done into nonvolatile memory with high write speeds such as phase change memory devices, but those are still in labs and years away from the market.

Re:Slow

[coolGuyZak]

Just hope you can fit enough usefull stuff in 4G, what with modern software bloat.

Considering that my current usage on my desktop fits wholly in 1GB of RAM, I don't believe there will be much of an issue there ;)

The same goes for my laptop (but w/ 512MB RAM). Caveat: These are both linux systems. Windows may use more (it's cache & page policies are rather different than linux), but I don't believe it is much worse than that.

Re:Slow

[big daddy kane]

Also, most flash drives you'll encounter use FAT32 as their filesystem. FAT32 is terribly inneffcient for flash memory and its awkward block sizes. To use FAT32 on flash memory, a Flash Translation Layer must be used, which allows the OS to see a FAT32 drive but to actually use a more effecient fs on the chip. Though this allows for greater compatibility, it does slow down the overall speed to the chip. If a flash optimized fs such as YAFFS or JFFS were used, performance and efficiency would be much greater.

Re:Slow

[linuxscrub]

Micron makes DRAM.

Ummm, mostly correct.

Micron also makes flash, but it is a small portion of their revenue.

In fact, Intel just partnered with them to increase Intel's access to flash product. The new company is called IM Flash:
http://www.micron.com/news/corporate/2005-11-21_na nd.html

scrub

Re:Slow

[Helios1182]

Linux is much more aggressive in RAM usage in my experience. Windows often leaves tons of unused memory free, while Linux tries to use most of it for caching of various things. I haven't seen either system use excessive paging under normal circumstances (assuming you have an appropriate amount of RAM for you usage).

Re:Slow

[chris_eineke]

Well, if so many people say that solid state sucks, it makes sense to think that. :P

Re:Slow

[Heembo]

As a guy who works on apps for Palm OS for a living

You poor soul - you still have a job? I though Palm OS was pretty much Dead? Whats the state of your industry?

Re:Slow

[adrianmonk]

You poor soul - you still have a job? I though Palm OS was pretty much Dead? Whats the state of your industry?

On the one hand, the rumors of its death have been greatly exaggerated. Palm has announced a Windows-based PDA/phone, and that wasn't exactly a vote of confidence, but they have basically said they will continue to make new Palm OS devices. And the idea that Palm OS is being dumped and they are switching to Linux is a misconception: they are replacing the proprietary kernel that Palm OS runs on top of right now with Linux, but apparently the upper levels will look very similar to how they look now. (It may not even be immediately obvious to the user that it is Linux inside.)

On the other hand, PDA sales in general have been declining as cell phones have duplicated most of their functionality. The two most useful things a PDA has are a calendar and a contacts list. Most cell phones have both of those now. And many cell phones can synchronize to your desktop via Bluetooth or USB. And nobody wants to carry TWO devices in their pocket (a PDA and a cell phone), and if they have to choose, guess which one loses? So, the Palm OS market is not really a very pretty picture unless Palm OS can break into the smartphone market, which so far it hasn't succeeded in doing very well.

To make matters worse, in a sense it doesn't really matter if Palm OS is dead. What matters is whether people believe Palm OS is dead. If they believe it is, they're not going to buy a Palm device. And then it becomes a self-fulfilling prophecy.

So yes, I still have a job, but it may not be a very good area to be in long term. Or even medium term. :-(

Re:Slow

[djupedal]

Micron also makes flash

Not correct - Micron does not manufacture any flash products.

Micron markets flash manufactured and supplied by another (JV manuf.) company...different from 'making' (manufacturing) it themselves (note using $$ from Apple).

Intel, Micron team up for NAND flash push 1. EE Times' top ten stories of 2005 Peter Clarke (12/16/2005 9:55 AM EST) URL: http://www.eetimes.com/showArticle.jhtml?articleID =175004196
LONDON -- This year saw the announcement of the plan to create yet another joint venture manufacturing company -- IM Flash Technologies LLC -- which will supply its parents, Intel and Micron with NAND type flash memory.

Re:Slow

[Heembo]

Aw man! I hate to see skilled programmers out of work. Have you started looking at your next major career move? Though of web-centric programming (Java, PHP, Ruby, .net), or are you thinking embedded is where you want to stay?

Re:Slow

[LinuxLuver]

Well...he did say in 5-6 years....so that must be about the time the fast solid state storage is expected to be a lot cheaper. Five to 6 years ago most PCs being sold were under 1GHz........

What will this do to OS requirements?

[sfcat]

With solid state memory, won't you never have to reboot the OS? Will I still have to reboot Windows every so often even though the machine is capable of instant on/off? This feature of the hardware will put serious reliability requirements on all OSes. MS will have to finally fix the damn blue screen or its lack of reliability will be a serious henderence.

Re:What will this do to OS requirements?

[TubeSteak]

My understanding of Vista is that you can reboot 'sections' of the OS without having to restart everything.

As for the blue screen, I hear MS fixed that in the Xbox360.
They made it black.

thank you, try the meat loaf and don't forget to tip your waitress.

Re:What will this do to OS requirements?

[Jeremi]

With solid state memory, won't you never have to reboot the OS?

There are plenty of people who never (or almost never) reboot their OS now... they just leave their machine on 24/7 possibly with an automatic sleep mode or perhaps without. I agree that having solid state memory would probably do away with "reboots" almost entirely, but that situation wouldn't be that much different from how it is now...

Re:What will this do to OS requirements?

[StCredZero]

It won't be enough. Why? Because The Internet is for Porn! (A hilarious video via Google Video)

Re:What will this do to OS requirements?

[SanityInAnarchy]

Nope, you're thinking of non-volatile RAM. This is nothing more revolutionary than a huge pile of flash keychains -- and about as fast to use. So yes, laptops with this will still need RAM.

But yes, IF we do eventually get non-volatile RAM that's as fast as volatile RAM and cheap enough to replace hard drives, we will have to do some rethinking of OS and software design.

Well, true, we won't HAVE to -- stuff does work off ramdisks and tmpfs, but those are still designed to go away when the computer shuts off. The programs which were smart enough to mmap everything and let the OS manage the caching will be the fastest for a long time. And we'll have to decide how to represent the new memory. It seems to me you'd have to call it RAM and come up with a way for the ramfs to persist, while still allowing "reboots", because the alternative is calling it a "disk" (setting aside some space for "RAM") -- yes, you could create huge swap spaces, but swapping to RAM is still slower than just using the whole space as RAM.

Unfortunately, I'm guessing that by the time everyone stops hopping up and down about how exciting the new tech is, we'll have it running unmodified Windows with so many hardware kludges to make it work that the system is slower and stupider for everybody else, too.

Re:What will this do to OS requirements?

[DogDude]

MS will have to finally fix the damn blue screen or its lack of reliability will be a serious henderence.

C'mon, trolls! The whole "blue screen of death" thing hasn't really been a problems since the Windows 95/98/ME days. Can't you trolls come up with something new? This FUD is old and tired.

Re:What will this do to OS requirements?

[vertinox]

With solid state memory, won't you never have to reboot the OS?

Although, I doubt it will be in his key note tomorrow for a new product release, I think Steve Jobs has a plan for something like this. I remember a few years back in either 2000 or 2001 that he discussed boot times and how they wanted to someday get that to 0 because people waste hours (if not days) per year waiting for their computers to boot.

Two reasons I see they are doing this is they know Windows Vista will support something and if they can get their version out the door first and secondly, if they have flash ram built into the system as part of the OS it will be harder to put the newer versions of OS X on non-Mac hardware mostly because the OS expects the flash memory to be there.

Besides, they've got buttloads of Flash to do this already from the Nanos. Chances are this will be in tandem of whatever Intel is doing with their hardware.

But I doubt we will see anything in stores until 2007.

Re:What will this do to OS requirements?

[Hachey]

That was the best lowbrow joke I have heard in a week. Kudos to you! :)

change is bad

[loserhead]

i dunno...i would rather use hard disks personally. in my ecxperience, they fail in a less catastrophic way. have a few errors....back it up and get a new HDD. with flash memory, when it fails, it FAILS. the end

Re:change is bad

[drgonzo59]

You obviously didn't own an IBM Deskstar and didn't have to hear the "click-o-death" and all your data gone before you can type "tar -cf" -- which is why they became known as Deathstar. If that is not catastrophic then I don't know what is...

I would love to see the HDs go, they are the only critical moving component in the system. Anything that is moving at 10,000+ RPM is prone to failure. I also wonder whatever happened to holographic solid state memory that was supposed to hold TBs of data - that stuff has also been "right around corner"... but then again, so has the time machine...

Re:change is bad

[E8086]

I'll go with both, I either read about a combo flash/hdd drive somewhere or said it would be a good idea somewhere. I also prefer the use of HDDs for storage of important data, but there's no reason flash can't also be used, for 'less important' data, say the partition containing the Windows and apps installations, around 4GB on my PC. I'd have no problem with having windows and apps on a fast 6-8GB flash drive that I can backup once a month/week/day or whenever I install new software or make any major changes. That way if the flash drive fails all you'd have to do is RMA it or get a new one, restore the backup to the new one and carry on as usual. I'm sure there are people who would risk total loss of data, as if they make backups, to have the boot time cut for minutes to seconds and have fifteen hours(so claims the article) of battery life per charge of their notebook.

Re:change is bad

[RussR42]

We should obviously focus on the time machine, as when it's completed we can make sure the holographic memory is/was right around the corner. Hell, perhaps I can send back some of that holographic memory already filled with pr0n and thus releave the load on the internet.

Do-dee-do la-de-da... Hmm... looks like it didn't work. damn. nevermind. Good thing azerus is already running.

Re:change is bad

[Rickler]

i dunno...i would rather use hard disks personally. in my ecxperience, they fail in a less catastrophic way. have a few errors....back it up and get a new HDD. with flash memory, when it fails, it FAILS. the end

But when is the last time your hard drive has been through the wash and still worked afterwards?

Re:change is bad

[Urusai]

The time machine is indeed right around the corner...tell me, do you know where all your socks are? Those time voyagers are such pranksters.

Re:change is bad

[Wyatt Earp]

There are tons of examples of bad engineering we can pull up as arguements for and against a technology.

But the Deathstars were years ago and generations ago for hard disks, I had a 14GB Deathstar...what, 7 years ago?

As for "anything that is moving at 10,000+ RPM is prone to failure", that's not true, jet turbines spin at very high speeds and are very reliable in the current generation models, something like .07 failures per 100,000 hours of operation. The speed something spins has no bearing on it's reliability, high speed systems can be very reliable, and most of your desktop hard disks are spinning at 5400 to 7200 RPM with laptops at 4200-5400 RPM

Re:change is bad

[Jeremi]

I also wonder whatever happened to holographic solid state memory that was supposed to hold TBs of data

I think the problem was that the capacity and cost of "traditional" hard drives has improved so rapidly over the past decade that holographic memory, like lots of other potentially revolutionary alternatives, simply didn't look like it would be able to compete any time in the forseeable future. Without a likely market for the product, research funding dried up.

Re:change is bad

[klang]

But when is the last time your hard drive has been through the wash and still worked afterwards?
My hard drive sits inside my portable computer and I usually notice it when it ends up with the dirty laundry... I even notice my iPod, so that's not an issue either..

Re:change is bad

[drgonzo59]

The point was that if the platters or the turbines weren't spinning at all, the reliability would have been a lot higher!

I was talking about reliability relative to the non-spinning components. Whouldn't you agree that within normal operating parameters (no power surges and good temp. control a video card or a processor would last a lot longer than a hard drive spinning at 10K RPM?

Intuitively: I had to replace two hard drives on my machine because of failure while all other components have been working without a glitch for 5+ years. The main difference is that hard drives contain fast moving parts (fans also move fast - so once in a while I have to check them or at least clean the dust away from them).

The same goes for the turbines - I would think they would have to replace the turbines a lot more often than they would have to replace rivets or bolts. Or to put it in other words if the jet turbine wasn't spinning it would have a much lower failure rate than .07/100K hours.

Re:change is bad

[Jesapoo]

Welcome to 1978...

http://en.wikipedia.org/wiki/Redundant_array_of_in dependent_disks

Re:change is bad

[despisethesun]

Intuitively: I had to replace two hard drives on my machine because of failure while all other components have been working without a glitch for 5+ years. The main difference is that hard drives contain fast moving parts (fans also move fast - so once in a while I have to check them or at least clean the dust away from them).

I, on the other hand, have just bought a new computer because the motherboard on the old one died. The computer was about 4 years old. Its hard drives are still going in the new machine without any trouble.

Re:change is bad

[Zaiff Urgulbunger]

But at least with a HD, if you really really really need to get your data back, then it is possible to employ the services of a data recover company to get it back. With something solid-state, I think it is more likely you will never be able to recover that data.

Not that a data recovery company is a good, cost effective thing, but at least there is a plan-b in the event you need one!

Re:change is bad

[drgonzo59]

That is true, I looked into that and the price was just way too high to justify saving my personal files...

Re:change is bad

[Wyatt Earp]

"I would think they would have to replace the turbines a lot more often than they would have to replace rivets or bolts."

No, jet turbines are one of the most reliable systems on an aircraft as long as they are maintained properly and are up to date jet turbines. The electronics packages are the most maintaince intensive systems on aircraft. The "rivets and bolts" aren't a system, that's like comparing the screws in your case to the hard disk or the cpu.

Nothing is free, so while we have spinning systems for the hard disks which can cause failures, what do we give up when we go to solid state storage?

Re:change is bad

[v1]

MTBF for hard drives, at least in my experience, is much shorter than for flash drives. I've owned five USB flash drives and four compactflash cards, as well as several dozen hard drives. Three of those drives have failed. Zero of the flash units have failed. At work I have seen dozens of hard drives fail, and not a single usb drive die. (though I did have to do data recovery on one flash drive with a corrupt partition) Nodding to another poster, yes, about 50% of those drives were deathstars. Half of the rest were quantum fireball 6.4gb. A few I was able to salvage data from, but drives like the chirping deathstars were a total loss. (if it won't spin, you're pretty much out of options) Not having any flash drives fail I can't say how easy they are to data recover, but I will admit I don't expect to have good odds.

There are data recovery businesses you can ship HDDs and flash drives to for recovery. Their odds of success seem to be better for flash drives, I think since a head crash is physical damage that has no real workaround. If you have enough money you can pay a lab to pop the top off the flash nand and stick it in a test fixture and pull the data straight off the storage arrays on the die.

Currently all flash drives are of a small enough capacity to be very easy and convenient to back up. I carry a 4gb cruzer, (large even by today's thumb drive standards) and it gets rsynced to my laptop about once a week, taking about 10 seconds. The laptop and its 80gb HDD get rsynced to the server about twice a month, and takes 10-30 minutes to complete. This makes the loss of a flash drive somewhat less serious of a problem for me at least. Less data to lose, and more recent backups available.

Comparing a hard drive to a flash drive, they appear to both have very similar structure. Both have an interface - the HDD has an onboard controller card for IDE interface and the flash has a USB interface chip. Both HDD and flash nand can experience "pinhole" failures, loss of a block or loss of a storage cell, and both have programming to compensate and "spare" the bad block so the error does not affect the user. A flash drive totally dying is no different than losing the onboard controller card on a HDD. (though in that case, admittedly, you might be able to scrounge up a replacement)

Hard drives are very vulnerable to physical shock while in use. I have seen warnings on almost all flash drives and CF cards I have bought that also warn against physical shock though I don't know why and have never heard of a flash drive failing due to impact. Twice I've seen a HD head-crash due to mishandling while spun up, and twice I've seen a HD require "impact maintenance" to get it to spin up.

I have a lot more confidence in flash storage than I do in magnetic storage. I also expect flash-based hard drives to be in at least 20% of new laptops five years from now.

When was this article conceived?

[Saven Marek]

The average notebook has 30GB (of hard drive storage). How long is it before the notebook has solid state memory?

I havent seen a laptop with less than 40GB in I dont know how long. A long time anyway. Maybe this is out of date.

Re:When was this article conceived?

[nukem996]

They could be founding old laptops as well. Infact more and more laptops are getting bigger hard drives(60gig and 80gig). Anyway they might be looking at what the average consumer uses. I know I only have 50gigs left on my 200gig drive after about 1.5 years of use.

Re:When was this article conceived?

[Wavicle]

Apple's most popular laptop, the iBook 12" only comes with a 20GB drive, and all of Dell's laptops start at 20GB too.

O RLY?

Re:When was this article conceived?

[Roydd McWilson]

I'm not sure what you're claiming. Both of those pages list the hard drive capacities for base configurations, not the "up to" capacities. I know as a fact that the current generation iBook G4s start off at 40GB.

Re:When was this article conceived?

[pintomp3]

maybe he wasn't refering to average new one. maybe the average notebook out there.

Re:When was this article conceived?

[Namegduf Live]

I don't agree with many of these comments, myself. I am sitting at a laptop with 20GB hard disk space, and working happily. The problem is that people aren't organised with their data. I have all my files for permanent storage under one directory - Archives. Within there are subfolders for Software (Installers), Music, Pictures, Games, Drivers, Guides, Patches, and such. It all fits on a DVD for being backed up. Now, if my music was less midi files and more mp3 files and more extensive, I can see it taking more space, but the typical music collection would surely be unlikely to hit 10GB, so a 30GB for someone with a more typical collection would be better than a 20GB for me. Another major part is uninstalling/deleting stuff when you're finished with it, and deleting built up junk in temporary folders and such. Otherwise, you are just going to fill it with junk you don't need over time, and any hard disk will end up full with that, eventually. It's just a matter of time. With those tricks, I've been multi-booting a recent version of Linux, with a GUI, and not one specially designed to be small, playing games (modern, multi-GB of HD space ones), browsing the internet, and so forth. Basically, I treat my hard disk as a place of limited volume, am careful with my use, and such, as opposed to filling it with junk and never uninstalling. A 30GB flash drive, if it goes down to a reasonable cost for a good speed one, would only be a good thing.

(This does not apply to those with extensive music libraries, who do video editting, or otherwise have need to fill their hard disks with stuff they actually use. I can understand that such uses would require significantly more space than this drive offers. I can see a day in fact, when for typical usage people have flash drives, and other drives are used for mass storage. Given flash drives are relatively mobile, the idea of carrying yours with you and plugging it into computers in internet cafes and hotel rooms as well as your own is an interesting possibility)

Re:When was this article conceived?

[Wavicle]

Didn't look to close at the pages did you?

As the previous post pointed out, both those pages were for minimal systems. Didn't you notice that one of those Dell's was only $499? It still came with a 40GB drive.

Re:When was this article conceived?

[Guspaz]

Older notebooks don't matter when considering next-gen notebooks. Any solid-state drive's size will be judged based on the magnetic drives available at that time.

Is this guy for real?

[barc0001]

"There will always be a need for storage, but when was the last time you tapped out a drive"

Last week at the parents' place. Two days ago at work. Probably tonight as well at home. You were saying?

No matter how much storage you put in a given system, it will eventually be not enough. I've seen it a million times.

Also, flash memory is way too slow to be used as primary storage. Putting 512MB of MP3s on my SD card takes almost a three minutes. Drive to drive, that's under 10 seconds.

And let's not even mention how quickly a cache partition would die with the 100,000 writes before failure standard of current flash drives...

Re:Is this guy for real?

[wyldeone]

Also, flash memory is way too slow to be used as primary storage. Putting 512MB of MP3s on my SD card takes almost a three minutes. Drive to drive, that's under 10 seconds.

That's totally wrong. The whole point of using memory instead of a hdd is because of speed; the long time for your mp3 player to fill is due to the transfer rate of whatever you're hooking it up to (ie usb).

Re:Is this guy for real?

[Anonymous Coward]

Also, flash memory is way too slow to be used as primary storage. Putting 512MB of MP3s on my SD card takes almost a three minutes. Drive to drive, that's under 10 seconds.

SD cards are slow. Flash memory is not.

Re:Is this guy for real?

[AEton]

"There will always be a need for storage, but when was the last time you tapped out a drive"

Last week at the parents' place. Two days ago at work. Probably tonight as well at home.

Most civilized people do not discuss this sort of thing in polite conversation.

Re:Is this guy for real?

[karnal]

The whole point of using memory instead of a hdd is because of speed; the long time for your mp3 player to fill is due to the transfer rate of whatever you're hooking it up to (ie usb).

That's not entirely correct.

While if you hook up a flash memory to the USB 1 spec, it will be painfully slow, even with a connection to a high-speed USB 2.0 hub, you'll still run into slowdowns. Why? Because most flash (which is most, if not all non-disk related MP3 players) write speeds are averaging around 5-10MB/sec. And even then, that's being generous.

So, for 10MB/sec, that would be at least 1 minute to fill up a 512MB mp3 player. Of course, real world is never the same as rated specs, so I'd be happy with 2 minutes, to be honest....

Another neat trick to try with Flash drives is to fill them with a bunch of itty bitty files - it literally takes forever to do so! Maybe someone more insightful than I can enlighten me as to why that is....

Re:Is this guy for real?

[CosmeticLobotamy]

No matter how much storage you put in a given system, it will eventually be not enough. I've seen it a million times.

I remember begging my mom to replace our 2MB hard drive with one of the fancy new 20MB ones. "But Mom! That's twenty MILLION letters! You'll NEVER use that much. You don't type that fast."

Then some jerk went and invented graphics. Bastard.

Re:Is this guy for real?

[AdriaanN]

The slowdown is due to the file system: FAT. Everytime a (small) file is written, the File Allocation Table needs to be be (re)written too.

It's just one of those great things MS has given us.

Re:Is this guy for real?

[barc0001]

SD cards are slow. Flash memory is not.

True, but have you got any interfaces that ARE fast for flash memory? Hard drives are quite speedy as well, it's just having to move that damn head around that makes them look bad, after all.

Re:Is this guy for real?

[imroy]

Another neat trick to try with Flash drives is to fill them with a bunch of itty bitty files - it literally takes forever to do so!

It literally takes forever does it? Then I take it you're still waiting for those itty bitty files to copy? Better hope you don't have a power outage...

Re:Is this guy for real?

[captaineo]

I'd like to know what kind of hard drives are giving you over 50MB/sec in real-world conditions!

Re:Is this guy for real?

[pintomp3]

at home, yes. with computers becoming the digital dumpster for pictures, audio and video you will never have enough storage. though that's assuming they are. my parents have a 80GB drive from last year. over 60gb of that is still free. and at work, most users don't work with multimedia. we only have 3-4 users (out of over 150) that have more than 5GB of files. this includes email. and speed, the fastest sd cards can read/write about 22MB/sec these days. they will only get bigger/faster/cheaper/smaller/etc. for the server or any other uber-storage needs, hard drive will always be the cheaper solution. but going solid state will allow computers to get smaller and quiter, much closer to the "appliance" people are looking for.

Re:Is this guy for real?

[akuma(x86)]

"30GB should be enough for anybody" -- say the CEO OF A MAJOR FLASH AND DRAM company.

Wow - I'm shocked!

People will always find ways to fill up what you give them.
How about redundant storage for reliability?
Effectively infinite undo/save for documents (Never worry about "saving" a document again because every keystroke is saved).

etc...

That's just out of my ass. I'm sure other people have many more usages for more space.

Re:Is this guy for real?

[baywulf]

"Another neat trick to try with Flash drives is to fill them with a bunch of itty bitty files - it literally takes forever to do so! Maybe someone more insightful than I can enlighten me as to why that is...."

The NAND memory used in flash drives are optimal for sequential writes due to the large erase blocks which can reach a couple hundres kilobytes. When you write small files, it has to copy everything in that erase block to a new location except the small portion it changed. This results in significant overhead. Reads don't have that issue since no modifications are done.

Re:Is this guy for real?

[cskrat]

Checking 512MB in 10s for you.

It took me 11 on my box going between two old 120gb maxtors on different channels.
Maybe the poster you're refering two was using newer SATA drives or just going by the seat of his pants and meant "about" 10s.

Many current hard disks can sustain 60-70 MB/s in benchmarks so 50 isn't too wild of a figure.

Re:Is this guy for real?

[AVGVSTVS]

Don't be an ass, he was using a figure of speech. It Literally takes forever and a day.

Re:Is this guy for real?

[Breakfast Pants]

First of all you can use multiple chips--say with eight chips you distribute each byte in storage as one bit on each chip, then the speed with which one chip can return a byte turns into the speed with which you can get 8 bytes. Second of all (in response to the 100,000 writes lifetime) there have been filesystems optimized for flash chips for a very long time--hint they don't write files to the same area when they are changed, they propogate new changes in a kind of cyclical manner, approximating the theoretical (number of bits)*100,000 amount of bits that you can write (the realistic number is much less given that writes tend to be in a concentratedly small number of files and there are a lot of files that just hang around unchanged, but as long as there is a decent amount of free space the 100,000 writes is really no limitation with these filesystems). Filesystems like this are used by people running linux on their ipaqs, etc. etc.
 
So the two main problems solid state storage has vs. harddrives are easily taken care of (throughput and lifetime). They already win out on latency, energy efficiency (particularly since HD's constantly keep spinning whether or not data is currently needed), reliability and noise--the only thing left is and price.

Re:Is this guy for real?

[shess]

While if you hook up a flash memory to the USB 1 spec, it will be painfully slow, even with a connection to a high-speed USB 2.0 hub, you'll still run into slowdowns. Why? Because most flash (which is most, if not all non-disk related MP3 players) write speeds are averaging around 5-10MB/sec. And even then, that's being generous.

Individual RAM chips are also slow. That's why they access them all in parallel, even though these days you can access multiple bits in parallel within a single chip. If you were going to use flash for system storage, you'd most likely pull the controller at least partially offchip, and start managing them as something similar to a RAID0 array. Writing 512 bytes will be slow, but writing 16x512 bytes won't be any slower, something like that.

-scott

Re:Is this guy for real?

[Alef]

No matter how much storage you put in a given system, it will eventually be not enough. I've seen it a million times.

I think data behaves similar to gases. It expands to fill whatever space it occupies, and when you compress it the computer gets hotter.

Re:Is this guy for real?

[Harik]

That's totally wrong. The whole point of using memory instead of a hdd is because of speed; the long time for your mp3 player to fill is due to the transfer rate of whatever you're hooking it up to (ie usb).

You're totally wrong on that. Flash memory is NOT SDRAM. Write speeds (per chip) are on the order of 128 kilobytes per second. Contrast that with harddrive speed, which are on the order of 20 megabytes per second. Even with faster flash and slower laptop drives, that's 100x speed difference. TWO orders of magnitude.

In summary, this guy is dreaming if he thinks anyone will buy flash harddrives in the next 5 years. Even high-end cameras went to microdrives rather then flash storage just to get the picture->picture time down to a reasonable level.

This has already begun...for desktops too!

[meatflower]

Gigabyte has something out they call i-RAM. It's a PCI add-in card that allows you to plug regular ram sticks into and then access them as a piece of solid storage space. They say its good for "multimedia applications" and I'm sure it is...if not a little overkill.
 
  Here's a link to a review from Anandtech http://anandtech.com/storage/showdoc.aspx?i=2480

Re:This has already begun...for desktops too!

[rolfwind]

I believe that's just a normal ramdrive - they've been around forever with software emulation. Of course, the advantage of a hardware add-on is that, otherwise, you have to part with a portion of your system ram to make it into a ramdrive and of course, it was not always economical to add more ram (limited # of slots or too expensive for a huge single stick of ram) - the PCI card effectively just doubles the number of ram slots you have.....

Cenatek also has a Solid State Disk hardware solution available for a long time:
http://www.cenatek.com/

But it's always been ridiculously expensive to me for what basically is ram on a PCI board.

A ramdisk is very nice for any files (or a shitload of small files) that get read and written to a lot but don't get loaded into system memory for some reason. I used to use a software one (emulated another drive as "G:\" back when I used Windows) and pointed my browser there for it's cache folder into it - speeded up the whole surfing experience with already visited webpages...

Re:This has already begun...for desktops too!

[Val314]

i have seen something like this RAM Drive back in the ISA-Bus days... it hasnt been quite successful

Re:This has already begun...for desktops too!

[Urusai]

Yeah I'm surprised nobody mentioned this inanity earlier. It would be cheaper (and faster) to ditch your current RAM DIMMs, upgrade to some fat >1G chips, and set up a RAM drive. This is why I hate the industry, anyone with a clue is just dead weight, we need fresh talent to reinvent the wheel over and over.

Re:This has already begun...for desktops too!

[vux984]

It would be cheaper (and faster) to ditch your current RAM DIMMs, upgrade to some fat >1G chips, and set up a RAM drive.

You might want to rethink that because it won't work:

1) Most editions of Windows only support 4GB of RAM in TOTAL. Including XP Pro, Server 2000 and Server 2003. The 95/98/ME line only supports 1GB of RAM. Its going to be pretty hard to dedicate 4GBs of RAM to a software RAM drive if that's all (or more) than your OS will recognize. (Only Enterprise editions of Windows servers will address more than 4GBs.) How many linux distros support more than 4GB of RAM right now "out of the box (ie from the live cd/dvds or precompiled isos)

2) Most desktop MOTHERBOARDS don't even support >1GB chips or more than 4GB total RAM, including 'gamer' oriented boards like the ASUS A8N32-SLI, for example. You aren't going to have a 4GB RAM drive if you can't put more than 4GBs onto the motherboard. Generally only expensive server boards support more than 4GBs.

The i-RAM lets you build a 4GB RAM Drive today, and add it onto your system *without* sacrificing any system RAM, without installing a new OS, without getting a new mobo. Plus you can max out your system RAM, and then add an i-RAM on top of that!

Anandtech kept saying they couldn't see why you'd use an i-RAM over adding more memory; and they are right... except that maxxing out your system RAM is actually pretty easy; and what do you do THEN? What if you've already got 4GBs of RAM and photoshop is still paging on you? You CAN'T just throw more system RAM at it. i-RAM technology could be a solution.

Finally, another major difference between an i-ram and a software ram drive is that you can't install and boot an OS from a RAM drive.

(PS I am not affiliated with gigabyte or i-ram in anyway.)

cheers

Re:This has already begun...for desktops too!

[wfberg]

I believe that's just a normal ramdrive - they've been around forever with software emulation.

Very observant. Except;

• It doesn't need software emulation, it's transparent to the OS, in fact, you can boot off of it.
  
• It has battery power back-up, so if the computer shuts off or the grid goes down, the data is retained.
  
• Seeing as it doesn't use emulation, even if the OS goes down for some other reason; data still there. You can even do without write-behind cache (seeing as the cache would only be in system DRAM anyway), so you never have dirty data to flush!
  
• The RAM used on the PCI card doesn't come from the systems's RAM, no need to worry about bios/OS/architecture memory limitations (4GB?).
  

These cards are intended as a hard drive replacement for very demanding applications; for example high-volume transactional systems. Transactional means you want persistence, even in the face of power-outages or OS failure, but high-volume means that you can get quite a boost if random access is nice and fast (near zero seektimes). If your whole database won't fit in a few GB (pretty likely) and you're not distributing this sort of thing, it would still be great for transaction logs, temporary databases, sessions, etc. Or how about using them for message queues? Any message sent is persisted, but not written to a slow hard drive or database.

NAND drives I'm not too sure about. But for demanding applications, battery-back-upped-DRAM-drives are way cool.

Re:This has already begun...for desktops too!

[hanssprudel]

I used to use a software one (emulated another drive as "G:\" back when I used Windows) and pointed my browser there for it's cache folder into it - speeded up the whole surfing experience with already visited webpages...

I have a better idea. Why don't you make a RAM drive to put you swap file/partition on! That will make swapping real fast!!!!

(HINT: Browsers only use the disk cache when they have filled up the memory cache...)

Re:This has already begun...for desktops too!

[origin2k]

For mission critical applications ECC memory would be a requirement which this product does not support. Thus I wouldn't trust this solution for my database logs or data.

Re:This has already begun...for desktops too!

[TheTwoBest]

Um....last I checked, both server 2000 & server 2003 can handle more then 4gb of ram. This might not be for the most basic standard edition, but advanced or enterprise (don't remember the names for different lines) can handle plenty more. The limitation is that no single process can use more then 4gb of memory (so I don't know how a ram disk on that would work.)

Additionally, any descent mobo these day can handle greater then 1gb and good server boards can handle over 4gb without a problem.

Re:This has already begun...for desktops too!

[rolfwind]

I have a better idea. Why don't you make a RAM drive to put you swap file/partition on! That will make swapping real fast!!!!

(HINT: Browsers only use the disk cache when they have filled up the memory cache...)

Because that would be filled up with memory for all my swapped programs, not just my browser. I only want certain programs speeded up, not the ones my OS is deciding to put in the background/swap-out - mostly because my software ramdrive was small to begin with.....

Re:This has already begun...for desktops too!

[Trepalium]

A couple nits. The standard versions of Windows 2000 and 2003 indeed can only handle 4GB of memory at maximum. The Windows 2000 Advanced server, or Windows 2003 Enterprise or Datacenter edition can handle up to 64GB of RAM via PAE. Each application is limited to 2GB on any edition, unless the /3GB switch is used for the kernel, in which case, the maximum addressable application space increases to 3GB, and the maximum RAM decreases to 16GB. An EMS-like system (for those who remember writing programs in DOS) exists for accessing RAM outside the limits of the normal 2GB or 3GB address space allowed for applications.

Re:This has already begun...for desktops too!

[wfberg]

Perhaps not this products, but the fine folks at texas memory systems will hook you up with that (be it for a price, but I couldn't be bothered to find the "#2 we try harder", TMS is the market leader for DB oriented solid state drives).

Wrong direction

[ThatGeek]

I think that they're moving in the wrong direction. Yes, solid state is cool (despite its price). Yes, it uses less power (but is noticeably slower). What I want to see as the future of portables is a thin client. Companies try to roll out thin client desktops every few years, but they never seem to think about thin client portables. Imagine a very small portable that is nothing but a thin client with wireless. It wouldn't take much power, could run resource hungry apps via an ssh tunnel to a real box and be and be relatively cheap to produce. Something like what I saw on one of the blogs at Sun a few days ago represents the future. Don't try to take the whole computer with you, just take a small phone to call your computer.

Re:Wrong direction

[TubeSteak]

That's way to cool to ever really happen.

It is not an actual or planned product. Images are obviously made with GIMP. This is my vision or prediction for 2006. Read it as my pet peeves about today's iPods or mobile phones - cant keep them in a wallet, limited storage, limited processing capability, need for a charger, monopolistic service providers who try to squeeze every pennies out of your pocket for silly things like incoming calls.... I am just tossing an idea, catch it and build it if you want :-)

Re:Wrong direction

[krysolid]

I am familiar with the current Sun thin client. The paradigm "seems" to make
sense, until you realize just how fast technology leaps. My conclusion is that
what you get with a thin client is:

  * Yots of your data flowing through eveyone's networks.
  * Your data residing on someone else's "thing" somewhere.
  * A regular fee that someone is charging you to do everything for you.

This makes sense, I don't say it doesn't. But for me, I would prefer to
pay the price of waiting for all of it to become available for me to use
at home.

If they had the thin client, I can have a laptop and more or less
control my data, and applications. It is work, and costs money, but
theoretically the market should take care of that ... except for the
monopoly in Redmond, it would have been.

So, if they can stop or co-opt open source, then there will be nothing
but the thin client. Imagine what the government will use as an excuse
to keep their control to have to look through all your data.

You will have more and more data, maybe in the future, every moment of
your life could be stored ... do you want the government or corporations
looking through that? Even if it is "for your own good"???

Re:Wrong direction

[Jeremi]

Don't reject the entire idea quite yet -- the "thing somewhere" could be your own PC back at your house (via the Internet), or it could even be your laptop/cell-phone/PDA across the room in your briefcase (which you use for more intensive stuff, but the thin client is more convenient for quick google searches, IM, etc)

There's nothing about this idea that says you have to let some corporation or government hold your data for you.

Re:Wrong direction

[vux984]

In fact a huge number of us already do this with Remote Desktop in XP, or VNC anywhere, or Timbuktu on Win/Mac .. etc

Plus its arguably more secure than having the files on a laptop that might be stolen next time you turn around. I'd generally rely on the the security of an encrypted vpn session and the fact that only keystrokes and images are being sent accross the wire over the potential for a laptop to be stolen and the files accessed.

New cellphone devices have wireless internet access, and terminal clients available to them... the future is now... well.. yesterday; because these have around for a couple years now.

Re:Wrong direction

[Registered Coward v2]

I think that they're moving in the wrong direction. Yes, solid state is cool (despite its price). Yes, it uses less power (but is noticeably slower). What I want to see as the future of portables is a thin client. Companies try to roll out thin client desktops every few years, but they never seem to think about thin client portables. Imagine a very small portable that is nothing but a thin client with wireless. It wouldn't take much power, could run resource hungry apps via an ssh tunnel to a real box and be and be relatively cheap to produce. Something like what I saw on one of the blogs at Sun a few days ago represents the future. Don't try to take the whole computer with you, just take a small phone to call your computer.

We used to call them mainframes and terminals - and were thrilled when laptops let us run apps away from the mainframe. Of course, there was this thing called a modem that actually connected a terminal at home to your mainframe, but data transfers were painfully slow.

Seriously, while it works for thethered systems it's a real pain unless you have a reliable high band width connection.

Re:Wrong direction

[dodobh]

Thin clients don't make sense for small businesses/home users. They make sense for larger businesses, because the cost of administration comes down.

I don't see thin clients replacing laptops as much as replacing desktops. The lack of moving parts, and complex operating systems is the real bonus. Thin clients work very well in the Unix and mainframe paradigms. They do not work as well in the windows world.

Re:Wrong direction

[vertinox]

It wouldn't take much power, could run resource hungry apps via an ssh tunnel to a real box and be and be relatively cheap to produce.

That would require a ubiquitous wireless internet connection everywhere on the planet or maybe FiOS rolled out to everyhome to get the extreme bandwidth needed.

I'm not saying it won't happen, but I'd give it a time frame of 5 years when internet is available on every square inch of the planet. (and that might be an optimistic figure)

Lifespan

[sincewhen]

What is the expected lifespan (in cycles) for flash memory? I thought it was only good for a few thousand writes.
Has it improved recently?

Re:Lifespan

[baryon351]

What is the expected lifespan (in cycles) for flash memory? I thought it was only good for a few thousand writes.
Has it improved recently?

This topic arose when people started using flash memory as a hard drive in old Powerbook 1400s. While they're a nice very expandable old powerbook, they have a RAM ceiling of 64MB. a G3/400 CPU expansion in them is one thing, but being limited to 64MB is a pain in the butt.

So popping a flash ram card in and using it as the virtual memory drive let PB1400 owners use 128, 256MB of virtual memory, running off the flash ram which was far quicker than the internal HD for swapping. Many people have also used these cards as the main boot drive so the whole OS boots from RAM, swaps to that same RAM, and gives mostly silent operation and saves on battery life. Critics of doing this noted the drives would last a month or two until suffering write death.

Systems running these cards have been seen working just fine for 3-4 years now. Write limits in the range of tens to low hundreds of thousands may not seem much, but in reality it's working quite well. Apparently part of this is that most newer flash ram drives are set up to attempt evenly distributed writes over cells, and not concentrate hundreds of writes one after another on the same cell

Re:Lifespan

[mako1138]

I've looked at a few datasheets, and they all say 100 thousand read-write cycles.

Looks like Micron NAND flash is the same.

I read something about getting > 100k cycles out of an EEPROM, so some applications must approach that limit.

Re:Lifespan

[adrianmonk]

What is the expected lifespan (in cycles) for flash memory? I thought it was only good for a few thousand writes.

As I understand it, this is basically a solved problem. For one thing, the number of write cycles before a location dies is actually more like 10,000 or 100,000.

But more importantly, apparently most all modern flash controllers automatically and transparently cycle writes through various parts of the flash, so that if you tell the controller to write to the same block several times in a row, behind the scenes it will write to different physical blocks in the flash each time. It somehow maintains a constantly-changing mapping between block addresses and actual, physical blocks.

This process is called "wear leveling", and by doing this, the problem of wearing out one particular location is mostly eliminated. In fact, given the number of cycles that each block can be written, and given the transfer rate of a flash device, it might take years and years of continuous writing 24 hours a day before you can wear it out. Eventually, it will wear out, but most hard drives wouldn't stand up well to that kind of activity either, so you're no worse off with flash.

This article has a section on wear-leveling.

Re:Lifespan

[melonman]

My understanding, based on CF, is that it's fine for systems that boot from CF, run in RAM and only write to CF when the configuration changes (eg Locustworld's Meshbox), but pretty failure-prone on systems that write logs and otherwise use the CF as workspace (eg IPCop).

It seems pretty obvious to me that sooner or later we will get rid of the remaining moving parts in electronic equipment, but I'm not sure the alternative technology is there yet. Cf the time it took between the first hype of flat screens and serious in-roads into the CRT market.

Re:Lifespan

[Anonymous Coward]

I have a slight problem with this in general and with your scenario specifically.

Many small systems that I have setup boot from flash card, run in RAM and only write to the flash card when the configuration changes. Logging was forced to ram disk or disabled completely. This was done deliberately to avoid some of the problems with write endurance. I have had small Linux routers/firewalls running like this for 2-3 years now, so it must work. Distributing the wear across the entire flash card in this scenario might work quite well to extend the life since most of the disk is read mostly with only small, infrequent writes. I must assume, however, that there is still a finite life and that the system is doomed to fail at some point. Whether this is measured in 10's of years or 100's of years, I don't know and would have to do some lengthy profiling of the software and calculations to estimate.

In your scenario (i.e. running flash as a swap device) the flash memory is written much more often. In addition, I must assume that most of the flash is being written to almost continuously. After all, since memory limitation is what you're trying to overcome, I must assume that a large part of the 128M swap is being used most of the time. In this case, distributing wear across the whole flash is unlikely to do any good.

This site has calcs for predicting wear-out on flash-based disks:
http://www.bitmicro.com/press_resources_flash_ssd_ db2.php

The results range from almost useless (1.71 years) to who cares (130 years - I'll be dead by then) for 512 byte sectors on a flash disk device.

I design microcontroller hardware for a living and have had to deal with write endurance on EEprom (the technology flash is based on) for many years now. Although write endurance has been going up (it used to be several 1000, now it ranges to 10^5 to 10^6) it is still a concern in many of my designs. For instance, on a data logger I worked on recently, I went to a battery-backed RAM design rather than flash based. The battery-backed RAM design was good for about 10 years (limited by battery shelf-life) and the flash-based design would theoretically have worn out after only 6 months.

Last time I tapped out a drive?

[Dino]

Most of my hard drives are already full. Let's review.

ShuttlePC Red Hat box: ~120G on a 200G hard drive. (old IDE controller) Full.
G4 Apple Mac, 3 hard drives totalling ~ 620G. Aproximately 60% full, and that's only because I recently added a hard drive.
PC Laptop, 80G hard drive. 25% full. And that's only because the hard drive was recently formatted and reimaged.
120G external hard drive. 75% full
27G external hard drive. Full
60G iPod. Full

So I'm a little shy of a terrabyte of active hard drive space. It would all be full if I didn't have multiple binders full of CD-Rs and DVDs.

But I guess not everyone regularly edits and encodes video on their computers, or routes their entire entertainment system through their computers.

I don't think hard drives will ever be big enough because data files will continue to grow as well. Solid State memory is and will always been a niche technology for areas that suite it best such as high reliability, small packages and extreme environments.

IMHO the market is already awash in solid-state storage microcomputers. They're called PDAs.

Re:Last time I tapped out a drive?

[waferhead]

The fellow that said 30GB was enough has NEVER edited video.

Re:Last time I tapped out a drive?

[CaptDeuce]

G4 Apple Mac, 3 hard drives totalling ~ 620G. Aproximately 60% full, and that's only because I recently added a hard drive.

Dude, you should try emptying the Trash.

Hold on ... 40% of 620GB ... means you have about almost 250GB free. And you're complaining?

But I guess not everyone regularly edits and encodes video on their computers, or routes their entire entertainment system through their computers.

Oh.

Never mind.

Re:Last time I tapped out a drive?

[Eightyford]

Solid State memory is and will always been a niche technology for areas that suite it best such as high reliability, small packages and extreme environments.

That's a little short sighted, don't you think?

30 gigs is nothing

[mrraven]

My crappy old 900 hmz ibook has 40 gigs and I have to hook up a 120 gig firewire drive to it just to hold my mp3s and various digital video caputures, thousands of pictures and graphics, etc. So the answer is I would "tap out" 30 gigs instantly and you can add another 30 gig on top of that. Anything short of 80 gigs is really pretty laughable by todays standards, talk about years from now when we'll see hd (or blu ray) dvds, 5 channel 24 bit music, etc, etc.

Power vs Weight

[queenb**ch]

Let's set aside price. Does anyone know what the power consumption for say 100GB of this vs a 100GB hard drive is? If price weren't a consideration, I might be willing to consider a slight drop in speed, if it meant that my batter would last say 5x as long.

2 cents,

Queen B

Re:Power vs Weight

[silverkniveshotmail.]

I know you were just tossing a number, but I don't think solid state storage would reduce your power consumption by anything in that neighborhood. The biggest use of power in a laptop IINM is the backlight.

Re:Power vs Weight

[canuck57]

Let's set aside price. Does anyone know what the power consumption for say 100GB of this vs a 100GB hard drive is? If price weren't a consideration, I might be willing to consider a slight drop in speed, if it meant that my batter would last say 5x as long.

Space would be your next issue, for 100GB flash disk would have to be smaller too. Power would not be much different. Neither would heat.

There are alot of hurtles for this prediction to come true. I figure in 15-50 years this may become main stream in desktops and full functioning portables. Although I can see it for hand held devices right now. Some of those hurtles will be catching up with hard drive density and supassing it, heat, power, write speed (although read is fast), write/rewrite lifetime and cost.

What about limited number of rewrites?

[geneing]

I remember reading that flash memory can only be rewritten only about 10K-1M times. It works Ok for USB memory sticks, but having a page file on a solid state disk would destroy it in no time.

Re:What about limited number of rewrites?

[Jeff DeMaagd]

I think what that means is that you would have to force the operating system to not page out. That shouldn't be a problem because flash memory isn't that much different in complexity than RAM in terms of gates, so by the time you can afford get a 32GB flash module, 32GB of RAM would be pretty affordable too.

Re:What about limited number of rewrites?

[blincoln]

If you went back in time and told me in 1995 that in 2006 I'd have 512MB to 2GB of RAM, but still have to use a page file, I would never have believed you.

I'm sure it's possible to have entertainment center-type PCs that don't use them, but things like graphics and video editing software always seem to outpace the limits of RAM.

Re:What about limited number of rewrites?

[baywulf]

The endurance is much better than 10K-1M because for one, error correction is used recover from the first few bits lost due to endurance failure and secondly, wear leveling is used where a new location is written everytime a sector is modified.

Where are my mod points when I need 'em???

[PaulBu]

Please...

YOU KNOW!!!!

"Noticeably Slower"

[Nova Express]

it uses less power (but is noticeably slower)

I'm curious: Why is Flash/NAND memory "noticeably slower" than a hard drive? I remember back when RAM drives were the rage at the ultra-high end because of blazing fast access times. It wouldn't surprise me that Flash is slower, but it would surprise me that it was so much slower as to be slower than a HD, which has seek and access overheads.

Are you sure it just wasn't the way it was accessed (i.e., a USB drive)? I would think that a NAND drive on an equivalent bus (say FireWire or SCSI) would beat an HD hands-down. What am I missing?

Re:"Noticeably Slower"

[Andy Dodd]

Flash memory has great read speeds and pays no seek penalties.

Flash memory has abysmal write speeds that are outstripped by any modern hard drive even under pretty heavy seeking.

Tapped Out

[JollyGoodChase]

There will always be a need for storage, but when was the last time you tapped out a drive?

I tapped out a vein once...

hand held/palm/etc

[E8086]

It already exists in the form of all the hand helds out there, only difference is the scale, few hundred MHz instead of 2+GHz and storage in a few hundred MB instead of GB. My Zaurus(206MHz, 32MB ram, ~1GB storage(SD card) is running faster than my old laptop 366MHz, 320MB ram 12GB HDD. It suspends/restores in 2s and reboots in 1-2min while the laptop only takes a few minutes more. It's not going to replace the power of the desktop used for gaming or A/V editing, but it will improve the small, light and portable laptop/notebook.

Flashback. 1986 all over again?

[pair-a-noyd]

What's new about this?
Back about 1985 or 86 I bought a NVRam card for my AT.
I *think* it was called a "BatRam" or "BatDisk" or something like that.
I also had one before that for my 8bit XT machine.
I no longer have the 8bit card but I dug up the 16bit AT card out
of my garage just now, it took me about 30 seconds to find it.
Here's what it looks like, (please be gentle on my bandwidth!)
http://www.systemrecycler.com/misc/dscn0773.jpg
and
http://www.systemrecycler.com/misc/dscn0774.jpg

At the time, this was revolutionary stuff. You could power down and
all your stuff was right where it was before. I think these things were
only about 2 or 4 megabytes (which was HUGE back then).
IIRC, I was using mine as a ram disk. I could put LOTS of programs
on 4 megs. This being in the day when most programs were still being written
to run on 64k IBM PC's.

Re:Flashback. 1986 all over again?

[jez9999]

Mind telling me what make/model of keyboard you have there? Looks like something nice and simple although it probably has a load of stupid annoying internet keys along the top...

Re:Flashback. 1986 all over again?

[Slashdiddly]

How much did you pay for it?

Great for low end users

[TRRosen]

While 30 GB is a thimble for the Slashdot crowd. I've worked with a lot of lowend users (grandma's , email only) who only use 5-10 GB. A solid state drive would be perfect for them...smaller,less power,more durable (at least mechanically). Those who don't store any multimedia (MP3s, Movies,Photos) wont ever use more than about 5GB (3 for OS,1 for apps and a gig left for a whole ton of recipes and emails). I on the otherhand have two full 200GB drives and need to add more.

30GB? You've got to be kidding.

[wyldeone]

I don't know what decade he's in, but it's hard to find even a budget laptop that has only 30gb of hdd space. By the time the price of flash has shrunk, the files people will be using will be that much bigger, as will the harddrives. Without some major advance in the manufacture of flash, it will always be significantly more expensive per-gigabyte than hdd.

And the argument that 30gb ought to be enough for anyways (sound familiar?) is a fallacy. As disk space grows, so does the size of content and programs.

Confused

[NixLuver]

"Most solid state memory is pretty darn slow"? May I respectfully ask, What are you smoking? What do you think is in your system (main RAM) and on your processor (cache)? There is no comparison between the worst case access time on a 10ns CAS2 DDR memory bus and an 8 ms seek time hard drive. The memory is *orders of magnitude* faster than the hdd. That's why modern operating systems do read and write cacheing of disk information.

Re:Confused

[timeOday]

Nobody is proposing replacing HDDs with DRAM. Obviously the parent was referring to Flash.

My question is why we can't make DRAM chips as fast as desired by simply adding more parallelism. With a HDD it's pretty obvious you can't have a dozen independent seeking heads. But with Flash, can't they divide the bank into as many subsets as desired, and access them in parallel? If not, why not?

Re:Confused

[mikis]

Um, sorry, but AFAIK system RAM and cache memory does NOT qualify as "solid state memory". When you turn off the power, all the data is gone, so it's not really useful as a long-term storrage.

Now, the same problem bothers me too: every kind of affordable "solid state" memory I've seen -- USB drives, varoious flash memory cards -- is by order of magnitude slower than hard disk, even though they contain no moving parts. And, all of them have limited number if read/write cycles.

So what kind of technology do they have in mind as replacement for hard drives? I guess bunch of DRAM chips with power source does not qualify as practical or affordable.

Re:Confused

[Linknoid]

Whatever happened to MRAM (magnetic RAM) that IBM was working on?

Re:Confused

[NixLuver]

"Um, sorry, but AFAIK system RAM and cache memory does NOT qualify as "solid state memory".

That's because you apparently don't understand the term "solid state".

It's ok, I'll wait whilst you go look it up... There, make sense now?

Re:Confused

[NixLuver]

Actaully there are quite a few products available that do that very thing - battery backed IDE-interface drives that allow drop-in DIMMs that are identical to one type of system ram or another. With lithium backup batteries, they're good for a long time of power-off-storage, and they can-and will- max out the bus capabilities of your interface, whereas a drive will only do that until the cache is exhausted. Thus, I didn't find it all all obvious that the parent was referring to flash; I would note that a couple of other posters said "I'm assuming you're speaking of flash".

Your question about DRAM is more complex, If you were referring to DRAM and not flash. In essense, modern systems *do* multiplex memory, but since memory access patterns are difficult to predict, it's hard to build a chip structure that will allow one to significantly improve access time across the memory usage spectrum. If you figure out how to do that, I'm sure Intel and AMD would LOVE to speak to you! :D Instead, they fight over fractions of a percent improvement in branch chain prediction so that the prefetch queue can be properly populated.

Death to Floppies!

[ScottCooperDotNet]

Floppies won't be dead until Microsoft allows for Flash memory storage driver installs during Windows installs. Vista has a new installer so I'm not sure if they've taken the step needed to finally put the Floppy to rest.

Meanwhile, I'll keep charging $10 per floppy disk in need of recovery. It is amazing to me that people will save their only copy of a critical file on cheap disposable media.

I fill up drives like Wimpy eats burgers

[mooncaine]

I work with digital video and audio. I filled up 3 160 GB drives this year with stuff I can't delete for years, and I'll have my new 200 GB FireWire drive filled up by April. Yeah, I keep too much, but I have a lot of really, really large files.

Come tell me when they finally come out with FW3200 10 PetaByte thumb drives -- I'm going to need a few of those.

Re:I fill up drives like Wimpy eats burgers

[Jeff DeMaagd]

Why aren't you buying larger drives? I realize there's this cost per capacity issue, such that 400 and 500GB drives are kind of expensive, 300GB drives are surprisingly cheap. If you get too many drives running, then you start to get heat and noise concerns.

Re:I fill up drives like Wimpy eats burgers

[wenchmagnet]

You might want to take a look at http://www-306.ibm.com/software/tivoli/products/st orage-mgr-space/

Or come up with your own scheme of migrating data to tape... if you've got stuff you cant delete for years, you probably need to back it up. So if you can migrate those files off on to two sets of tapes, you're safe and have a scalable solution.

Re:I fill up drives like Wimpy eats burgers

[forkazoo]

I know what you mean -- I've got about a half-terabyte on m main desktop now, and I *really* need to move that to a terabyte because I'm nearly stuffed to the gills.

Re:I fill up drives like Wimpy eats burgers

[Slashdiddly]

I work with digital video and audio

Heh-heh, don't we all?

*checks on bittorrent progress*

Re:I fill up drives like Wimpy eats burgers

[mooncaine]

Good question. The last purchase I made was for 2 of the next-to-largest drive I could get, because I prefer modularity. That way, if one drive fails [or I drop it or spill something on it], I've still got the data on the other. Saves me time.

I don't run them all at the same time. I mount a drive for certain projects, and exchange it for another when I'm done. I've got two attached right now and one on the shelf.

Re:I fill up drives like Wimpy eats burgers

[mooncaine]

Hmm, you've given me an idea. Thanks, magnet. I work for a university -- there may be a way to get the computer center folks to make a tape backup for me [though when it comes to Macs, it's a hassle getting help in general]. I never seriously looked into the costs of tape backup. I should.

I'm there...

[Shanep]

if they can build lots of storage (50G+), lifetime you'd expect from DRAM (get rid of the wear problems of flash) and make it fast enough to saturate typical disk busses and fit it into 2.5" and 3.5" HDD form factor with appropriate bus to use as a HDD. Who needs TCQ when you have no head movement. Then these super fast disk busses would actually be worthwhile.

I've seen flash "disks" in HDD form factors for IDE and SCSI before, but they're hyper expensive, typically intended for military use and still suffer from the wear problems of flash memory.

I would love to see the end of mechanical disk storage. Somehow I imagine I would be wondering how I coped before it, like I do with the optical mouse. I wait in great anticipation! The power savings and added battery life should be great too.

Re:I'm there...

[Firehed]

Well aren't we all. You can get a DRAM-based 4GB drive that will most certainly saturate it's SATA bus, but it'll run at least $400 right now (Gigabyte i-Ram plus four 1-gig sticks; the former still seems quite unavailable). But zero seek times (well, a few nanoseconds) and making use of all the bandwidth available, with the durability of DRAM, and the only problem is the 12-hr battery life (though, that's only when the machine is unplugged, as it gets power from PCI which can and does tap the standby lines).

Assuming SATA stays on on the same track beyond SATA-III, we would have 2.4GB/s bandwidth in 2014 (bandwidth doubles every three years; SATA-III is slated for '08 at 600MB/s) which would just be approaching what you can get out of today's slower RAM. Now AFAIK there's no plans beyond SATA-III, and it's almost certain that eight years out, we'll have much faster solid-state storage dominating the market, probably with a $/GB ratio we've never even though possible, but the idea sounds delicious. As it is, the i-Ram is the only semi-sensible option, but it's availability is nil last I checked, and of the one place I found with a price quote, it was about three times what Gigabyte said it would cost (about fifty bucks), and you still have to buy the ram. Still, it'd make for a nice OS drive. Unfortunately, you can't install much more than that on it, and it's the other apps that tend to slow things down. Testing has shown that while it helps things feel more responsive, it's not going to be too useful until you can install a decent amount of software on it.

Re:I'm there...

[Shanep]

You can get a DRAM-based 4GB drive that will most certainly saturate it's SATA bus, but it'll run at least $400 right now (Gigabyte i-Ram plus four 1-gig sticks; the former still seems quite unavailable). But zero seek times (well, a few nanoseconds) and making use of all the bandwidth available, with the durability of DRAM, and the only problem is the 12-hr battery life (though, that's only when the machine is unplugged, as it gets power from PCI which can and does tap the standby lines).

I was quite excited when I heard about the i-Ram, however I soon realised that since it is SATA 150MB/s based and thus limited to that speed, it had lost most of its appeal to me. I would rather just buy that much RAM and use it with a software RAM disk to get the much greater speed that the RAM can provide as system RAM. I tested this out with various BSD's on an AMD 2800+ with 2GB DDR333 RAM, using 1GB for a RAM disk and I was getting over 700MB/s.

Sure the i-Ram is battery backed, but I'd rather have a PC with a UPS and then copy between software ramdisk and disk when the system comes up and goes down. I realise this is going to cause a higher risk, since the OS or ramdisk driver can crash, loosing the data, but exteme speed comes with the risks that avoiding constant commits to disk already has.

Maybe I'm being harsh, but I feel that this product can be in many circumstances too little too late. Its claim to fame is speed, but if extreme speed is desired, you can get Opteron motherboards which take LOTS of memory (I have seen 64GB with a single Opteron) and I would expect a big RAM disk on one of those could be pretty damn fast.

Or, you could avoid a software RAM disk in many circumstances and just exploit FreeBSD or NetBSD's fantastic unified buffer cache and go even faster. I found this to be the case (slightly).

All in all, I think 4GB of RAM would be put to much better use as system RAM. If low latency is what you're after, same deal. I'll take system RAM there also.

great for a laptop

[mrcdeckard]

30G is a bit thin, but think of this: having a 10Gig solid state partition for the OS, pagefile, applications and etc., then a 100G(or whatever) for all the data (mp3's, video, photos, if you have a lot). this way, unless you're really crunching data, watching a video or something, the OS would run off the RAM disk, which would save battery life, and, with possibly faster access times.

i might try this once usb sticks get up to 6G or more -- i only need 3G for OS X.

mr c

Mooreon's Law

[nick_davison]

He was quoting Mooreon's Law...

1. Every eighteen months, the technology that you support will double in capacity.

2. Every eighteen months, the technology that you are supporting it over will do nothing.

Ergo, given that average notebook hard drives are currently around 60gb, rather than 30gb, Moore's Law (as opposed to Mooreon's Law) allows us to deduce he began applying Mooreon's Law 18 months ago - the doubling of average disk space since then has been ignored by him as it's a competing technology (and covered under part II of Mooreon's Law) - it has just taken him this long to get anyone to take him seriously.

Applying Mooreon's law from that point we can deduce:

1. 1GB flash 18 months ago translates to 2GB today, translates to 4GB in 18 months, 8GB in 3 years, 16GB in 4.5 years, 32GB in 6 years.

2. 30GB of traditional 2.5 inch HDD 18 months ago should still be 30GB today and will be 30GB in 6 years.

Therefore, 32GB > 30GB in six years - hence 5-6 years is an accurate figure for when flash should overtake 2.5inch HDDs.

If, you know, Mooreon's Law wasn't for Mooreons.

Yeah no kidding, heard about HD

[SmallFurryCreature]

I recently succumbed and downloaded the HD version of a music clip (yeah I am a pirate, keel haul me you landlubber) and wow it looked good. It was a bit of a shock to find that it didn't actually fit my monitor but even scaled down it looked yummy. Only problem? about 12 minutes takes up 2 gb. Ouch

And one nice thing about laptops is that they come with widescreen pre-installed. All of sudden that 30gb drive doesn't look all that big.

It has always been the problem with solid state memory. The moment they increase the storage by 2 HD companies increase by ten and half the price.

Sure HD's are vulnerable, they may even suck your battery dry faster then an Itanium BUT they make up for it by being bloody big.

or to answer the question in the post. When did you last tap out a drive? Yesterday and I bought 2 sata drives each 320gb. Granted in part to retire an old Promise raid (yeah yeah I know, stupid) so I can finally upgrade my linux server to 2.6 but also because I maxed out all my disks.

Granted I wouldn't use a laptop as a fileserver BUT I am not going to accept a laptop with less storage then my bloody iPod. Not unless there are some very real advantages and multiple of them.

Far cheaper, far less power consumption, an awfull lot faster. I need at least 2 to make up for a huge loss in data storage. Oh and it better not be like flash wich craps out after using it to much. Geez learned that the hardway when I used my flash mp3 player as a backup solution.

Re:Yeah no kidding, heard about HD

[shmlco]

Yeah, I'm afraid that Mr. Appleton is engaged in a major bit of wishful thinking. I'm still chafing under the restriction that the largest available notebook drive is ONLY 120GB. I have a 11MP digital camera, and I can fill up a dozen gigabytes in a single afternoon.

Now, maybe if he makes an inexpensive 30GB USB thumb drive...

Re:Yeah no kidding, heard about HD

[Achromatic1978]

That's what I have, on both counts. Then I take it home to a 300GB Maxtor 'network storage drive' with another 400GB drive chained to it, which gives me a nice 700GB NTFS volume to mount. Problem (temporarily) solved.

Re:Yeah no kidding, heard about HD

[Znork]

"Not unless there are some very real advantages and multiple of them."

Frankly, the only way I can see it working would be as a modular system. Sure, if I could have a 'laptop system core' the size of a PDA with similar power consumption, with only enough storage for OS, basic apps and data, which could easily and reliably be extended with other storage depending on occasion. Self-contained bluetooth/wifi enabled disks, with their own batteries? Personal-space attached storage?

It's getting high time for a serious shakeup in the whole pda/laptop/phone/camera/media player field. I want a plug'n'play system, not carry around five devices with overlapping functionality that suck in different ways.

Re:Yeah no kidding, heard about HD

[GuyverDH]

Have you ever messed around with Linksys's NSLU2?

It's a great little gadget, with loads of promise.
I use it to hold ghost images of my windows systems, and digital archives.
The latest opensource firmware's unlock this neat little slug to allow it to do so much more than just disk sharing.

Oh - first thing you should do when you get it, (as long as you don't care about warranty that is), is de-underclock it back to it's fully rated speed.

By default, Linksys ships it with the speed cut in half, literally.

This is the way

[Delifisek]

Yea yea I know you can use Torrent and that donkey and of course you can easly fill that terabyte with dumping your divx pr0n collection.

And industry changin his way to solid state.

A 10 gig solid state very nice.

I see lots of people considers buying laptop ever before. Also that Mobile Phone sales was very high. That devices need large storage areas.

Solid State is absolute way.

And I wondering what about that IBM Holo Storage tech.

Re:human implants

[Lord Kano]

I'd settle for an interface. So that it's possible to upgrade the hardware without more surgery.

Can you imagine how much it would suck to be in competition with someone who had an implant that was two years newer than yours?

LK

Hybrid Designs are the Right Choice

[billstewart]

The right design for most applications is probably to use Flash for read-mostly drives containing your OS and most common applications, so they load fast without annoyances like rotational latency, and mechanical disk drives for bigger file storage (especially your media collection which is going to soak up whatever space you've got available), and enough caching RAM that you don't need to spin the disk too often. It's easy to do crude versions of this today (e.g. booting from Knoppix-in-flash and using power-management to spin down disks when you're not using them), and longer-term operating systems should be designed to take better advantage of flash to minimize startup latency.

Also, there are flash technologies that are starting to get rid of the limited numbers of write cycles that plagued earlier flash, as well as load-leveler drivers that even out the cycles on current flash.

Wow....

[Ichigo Kurosaki]

clearly the author hasnt gotten on any college networks recently....

MS Outlook Email Bloatware

[billstewart]

MS Outlook encourages you to use bloated email formats, e.g. mailing around Word documents and Powerpoints instead of text, and using various formatted text formats instead of plaintext.

I'm currently using about 1.5GB for the last year of work email, not counting the stuff I've deleted after saving the attachments or reading the contents. It'd be a lot bigger, except that Outlook apparently freaks out and dies if your .pst is 2GB or larger, and it also gets way too slow to search for things, so I try to keep it down to 500 MB (also made it easier to back up onto CD, back when I used CDs.) And I keep the older stuff in separate mailfiles as well. Not hard to burn 5-10GB that way, especially when people mail you lots of powerpoint stuff that you need to keep. My personal email, which Eudora mostly stores in Unix-mbox text formats, is an order of magnitude smaller for a similar number of messages.

But yes, it's a lot less than the space I burn on the machine where I download music :-)

4gb for the OS.

[Nik Picker]

and magnetic disk and removable media for the content ?

I must spend forever trying to configure windows to run the Documents and Settings folder on a seperate partition as well as finding ways of locking the current run state install for the Window folder. Programs like drive snapshot help in the build but I can see how it may now be possible to sell a alternative Pc.

Manufacturers can start builing a PC where you buy a seperate "option" of a external USB drive. Now they can manufacture a PC with all Solid State components from Motherboard, Graphic card, Memory and "disk" drive. Configure the operating system to boot and run and then allow it to locate other storage devices. Less power is then required for the overall build....

hey wait a minute .. isnt this Apples plan for the Mac Mini ? hmmmmm

Re:4gb for the OS.

[oojah]

My D&S is I:\Users\ - this is with 2000.

Searching through the registry, these bits look relevant:

HKCU\Software\Microsoft\Windows\CurrentVersion\Exp lorer\Shell Folders
HKLM\Software\Microsoft\Windows\CurrentVersion\Exp lorer\Shell Folders
HKLM\Software\Microsoft\Windows NT\CurrentVersion\ProfileList

Lots of applications will still use the old paths though. It's best to make the change first thing when you install Windows.

Cheers,

Roger

Re:4gb for the OS.

[Achromatic1978]

A better way to do it is this: Do as you said to move the D&S folder. Empty and recreate it, and then mount a partition in Disk Management at that point, using Change Drive Letter and Path > Mount in the following empty NTFS folder.

Then it'll appear as c:\docs and settings and work transparently from the other drive.

Re:human implants

[Lord Kano]

I try not to go beyond two years between major upgrades.

A week and a half ago I went from an Athlon XP 2800 to an Athlon 64 3000. My old CPU went into my GF's machine and that processor(which also was once mine) will be going into a machine for the children.

If we had implants, it would add a whole new level of ickyness to our current upgrade pattern. That's one of the reasons why I'm thinking that implanting an interface would be the better way to go.

LK

Hybrids?

[RickPartin]

I think what we'll be seeing is hard-drive/flash hybrids. Flash to store the OS on for quick energy conserving access and the hard drive for everything else. Flash just isn't ready to handle large amounts of data and it has a limited ammount of writes before it goes bad.

WinNT.sif and seperate partitions?

[ScottCooperDotNet]

I must spend forever trying to configure windows to run the Documents and Settings folder on a seperate partition as well as finding ways of locking the current run state install for the Window folder. Programs like drive snapshot help in the build but I can see how it may now be possible to sell a alternative PC.

I've read it is possible to the Unattended Answer File (WinNT.sif) to put the Profiles (D&S) folder on a separate partition in Windows XP (and likely 2000) but I've not been able to find a link.

Unfortunately not

[mustafap]

>won't you never have to reboot the OS?

Unlikely with Flash memory as your storage medium.

Although flash is non-volatile (it maintains its data when the supply to it has been removed, ) flash has a very finite number of times that you can write to it. Although lifetimes of 1,000,000 write and more are common now, that is still not enough for flash to replace main volatile memory. And of course the write time, although good compared to a harddisk, is very poor compared to RAM.

I totally agree with your sentiments though. Think of the power consumption we would save! SETI wouldn't like it though...

This is the year of solid state drives

[Anonymous Coward]

Everyone keeps mentioning their personal/private computers. Solid state memory will be big in corporate desktops. I'm a system administrator and where I work most of the computers use less than 2 GB. That's because only Windows and Office goes onto the drive. Very few additional programs are installed and documents are stored on a network mapped drive. This is what it's like at most of the larger workplaces.

I and my users would love to swap those 40-80 GB harddrives for 2 GB solid state drives and enjoy the benefits of a computer using less physical space, making less noise, consuming less energy, being faster and cheaper.

While it's generally accepted that harddrive space is cheap there is a minimum cost to pay. The smallest drives of today are still as expensive as the smallest drives of five years ago. With solid state drives I expect the price of the smallest drive to go away because it will be integrated on the mainboard - something that wull never happen with 3.5" drives.

See it in action now

[Tomaaah]

It is the first... ...personal computer configurable WITHOUT a System Hard Drive with all operating system and program files permanently sitting a on a Flash RamDisk. It operates at the speed of the maximum available memory bandwidth up to 8.5GB/s and is capable of over 150,000 I/O requests per second, with an average of 0.0% CPU utilization.

go-l.com have been selling desktops built on solid state ram disks for at least a couple of years from this site. That's the lowest model too.

Also their laptops included some of the same technology.

Oh, and checkout the monitors... woah.

AND.. no, I don't work for them.

Re:See it in action now

[Mr. Jerar]

You should do a search on litl.pk. It's interesting how the companies stock is showing up on the spam contaminated lists, so I would be careful with the website.

Re:See it in action now

[Tomaaah]

Interesting. They used to have a store a few years back. The site went down for a while and then came back. It didn't occur to me that perhaps the site as is has nohing to do with the company that actually was.

Re:See it in action now

[Mr. Jerar]

It's a shame to even have to suspect the site might not be legit but if it is legit then I like the way the company is heading with the products.

My Laptop is already Flash

[DrSkwid]

I got a 2.5" IDE<>flash adapter and installed a 1GB cf with miniBSD

It was 512 kB !

[wtarreau]

I think these things were only about 2 or 4 megabytes (which was HUGE back then).

It was a 512 kB (16*32 kB), but still was huge, and must have costed you a fortune back then ! I remember having made an NVRAM of 2 kB myself which I used to store some alternate boot code, and I was dreaming about the 6264 SRAM chips which would have brought my card to 8 kB ! It was back when you could modify your PC with a soldering iron.

Willy

same

[djupedal]

Based on Latin constare 'stand firm, stand at a price.'

cost (to the consumer)...same as price (paid)

cost
noun
the COST of the equipment PRICE, asking price, market price, selling price, unit price, fee, tariff, fare, toll, levy, charge, rental; value, valuation, quotation, rate, worth; informal humorous damage.

verb
the chair costs $186 be priced at, sell for, be valued at, fetch, come to, amount to; the proposal has not yet been costed put a price on, price, value, put a value on, put a figure on.

price
noun
the purchase price COST, charge, fee, fare, levy, amount, sum; outlay, expense, expenditure; valuation, quotation, estimate, asking price; informal humorous damage.

Differences in architecture

[MoogMan]

This is going to call for a change in the current architecture of systems. It's basically rendering RAM useless. It'd be cool if laptops with these RAM drives would basically scrap the RAM and increase the faster cache somewhat. Scrapping the main RAM should offset some of the price of these RAM drives, at the very least.

But, the main change is going to have to be with the software. Would there be any need for a swap file? Any need for memory scheduling? Wouldn't all programs be runnable from their location on the RAM disk, with no (or minimal) performance hit?

Your operating system wouldn't need to be "loaded", as such, as all the runtime information would be saved "in-place", as it were. We'd finally get "instant-on" systems, and hibernation would be a thing of the past.

Investing in flash memory makers

[Jackal2005]

Analysis of Flash memory, market, price point decline, flash capacity increase and companies are analyzed at: http://finnews.blogspot.com/2006/01/flash-memory-a nd-investing-in-flash.html

Topped out a 30GB harddisk?

[el_womble]

I think this is really interesting.

The role of the home desktop is changing. It used to be the powerhouse. The computer you used when you really wanted to get some work done... but that came at a price: working in an office. Laptops work for me, because when faced with a block the best way of solving it is a change of scenery. Sitting in the same place for hours on end for "fun" is less appealing now I have to do it at work as well.

My G5 is easily twice as powerful as my G4 Powerbook, but I use my laptop 80% of the time. So why have a the G5? It's a home server. I have over 40GB of music, 10GB of photos, 100GB of home movies and PVR, and its incredibly useful to have a single point of access for the whole household, and because its a desktop its always in the same place, always on and permanently connected to the internet meaning that not only does it server the house, it serves us whilst we're on the move as well.

Even if my laptop could match the desktop for storage, I wouldn't want it to be bogged down with running the services, and all the laptops in the house having independant media store is just plain bad management. Also, tasks like media recompression, code compilation and games are still done best on a machine with more RAM than sense and a processor thats designed for performance not low power consumption: you use a push bike to get to work and for fun, you use a car to do the shopping. Sometimes you need the heavy lifting.

In fact I now have a couple of home servers, but thats because I'm a nerd: I have a PIII running debian to provide the low power services like a front end for Azuereus, a few small web apps and LAN facing NFS server. Which is why I can't wait for a 20GB NAND drive that improves the battery life of my laptops. I just don't need that much storage on teh move providing I've got a decent wireless network connection.

As for, when was the last time I topped out a hardisk... yesterday. I hve 300GB of storage available to me and I use all of it. You can never have too much storage, you just don't need all of it, all of the time, providing you can access it from anywhere in the world network latency and speed is more of a barrier than local storage.

Re:Topped out a 30GB harddisk?

[ephex]

To take this one step further, it would be extremely nice to have the OS and some commonly used files available on the NAND drive and another hard drive available inside the notebook. Most of the time I would be able to run with the hard drive unmounted and powered off, saving a nice amount of power, and when I needed access to those files OS X could spin up the hard drive and mount it with no problem.

slight amendment necassary

[way2trivial]

I am after all, concerned with the trivial..

xp Pro 64 bit.. can indeed adress a little bit more than 4gb of ram..
(can't install jack shit for external hardware, but still)

Deja Vu (again)

[Anne Thwacks]

How long is it before the notebook has solid state memory? Five or six years,

I think I have heard this story ever January since 1970, and it was probably around before that.

A brief revue of the literature will reveal that, although its perefectly true that solid state memory follows More's law. HDs appear to as well.

At the time Bill Gates said "640k should be enough for anyone", a 40MB HD was the size of a Bendix washing machine, and cost about the same as a Ford Galaxie 500 with all the extras. 64k of RAM cost about ten times as much as a PC with no RAM.

In 1974, (check your library for old copies of Dr Dobbs) there was a serious debate as to whether the laws of physics made it impossible for memory to EVER cost less than 1c per bit!

And for those of you stupid enough to think solid sate means slow - ask someone what Google store their data on! People who know nothing about history are condemned to repeat it. The rest of us get shiney new USB thumb drives.

Dad, can i have a computer for high school

[cinnamon colbert]

sure honey, you can have my old laptop

DAAAAD it only has a terbyte of memory are you trying to embarass me
I hate you

OK, sugar pumpkin, what would you like

well the new laptops at CompUsa weigh 12 ounces, have a 3D screen, virtual keyboard, and a pB of ram (1 pB = 1^3 tB = 1^6 gB)but shirley jones has one
with an extra 20 pB so she can have evry movie ever made, and virtual language translation

Well, you know the jones have a lot more money then us, but how much is the pB machine...

And you cn just imagine what our great grandkids will say

In 5 years...

[wondercool]

.. hard disks have a capacity of at least 2 Tera bytes

30 GB looks cool in 5 years but everybody forgets that disks will grow as well.

So forget about solid state.

Re:In 5 years...

[klang]

solid state has uses.. as temporary storage in cellphones, digital cameras and mp3 players.

Storage capacity goes up, what you put on storage goes up in quality.

We need the 30 GB flash card in 5 years .. for that 25 giga pixel digital camera / DV camera that our cellphone will pack.

At home, we will put those pictures on the 2Tera byte harddisk .. live will not change much in 5 years..

Re:In 5 years...

[RexRhino]

Or, we could have a solid state drive for the OS (for fast bootup), and for software that might need to read/write lots of data to and from disk (such as doing manipulation on a video file too big to fit in memory)... and we could have a disk drive for everything else.

Not with flash

[Andy Dodd]

Flash memory has slooooooooooooooow write speeds. Far slower than the hard drives' read speeds.

Battery-backed DRAM or SRAM would be a whole other story... Even slow DRAM that is a few generations behind current technology would provide an insane performance boost if used as a drive cache.

Re:Not with flash

[adrianmonk]

Flash memory has slooooooooooooooow write speeds. Far slower than the hard drives' read speeds.

That's a good point, but I think there's a way around it. Simply put, on a laptop, at any given time, the hard drive is either spun up or it's not. So we have two cases:

1. The hard drive is spun up. In that case, you can just bypass the flash and write straight to the hard drive. Performance is never worse than with just a hard drive, because you can always adopt a write-around policy for the cache. (For what it's worth, ideally your I/O architecture would keep the write buffers around in RAM and then try to write those to the flash cache when the flash is idle, since anything you recently wrote probably ought to be in the cache.)
2. The hard drive is not spun up. In that case, you are fairly well screwed on writes without the flash, because even if you only want to write one byte to the hard drive, you are still blocked until it spins up, which could be several seconds. When you add flash as a write cache, you can start writing immediately. Even if the flash could only write at 1 MB/s, that's still a win because while the hard drive is spun down, the alternative is to be writing at 0 MB/s. And anyway, it seems like modern flash can handle writes much more quickly than that, on the order of 15 MB/s.

However, I do have to admit there is one wrinkle in this idea: the whole purpose of the flash is to be able to keep the hard drive spun down most of the time. Therefore, with this scheme, you will be more likely to need to have the hard drive spin up, because the odds of needing to spin up the hard drive on any given I/O increase as you increase the percentage of time that the hard drive is spun down. So, in situations where you have a cache miss, you will pay a big penalty, and assuming your flash really is slower at writes than the drive, you will pay a bit of a penalty (a smaller one) when your flash can't keep up with the writes and you have to spin up the hard drive to keep pace.

Still, I think for many workloads, the speeds could be quite good, and they'd be better with flash cache in some cases, and it could definitely result in power saving.

Re:Not with flash

[Andy Dodd]

Hmm, I wasn't thinking of using the flash as a writeback cache, I assumed you were talking about using it as a read cache. (Since overall, more reading is done than writing with most usage patterns.) The problem with using flash as a read cache from the hard drive is that you have to write something to the flash in order to read from it, and the flash memory's low write speed will be a huge bottleneck here. It really doesn't work that well if you use something as a cache that has a lower write speed than the read speed of whatever you're caching. :)

Volume-priced PC100 SDRAM please! IIRC, PC100 had a 100 MHz clock and was 32 bits wide, giving around 400 MB/sec of transfer rate, 8-10 times the sustained read speed of most hard drives. 4 gigs of that as a drive buffer would be great.

Hmm, I wonder if low-speed SDRAM chips are available at reasonable prices... The sticks are obscenely expensive per megabyte, but that's probably because of price-gouging the few people who need PC100 sticks. The ICs themselves not in PC100 format might be much cheaper.

Re:Not with flash

[adrianmonk]

Hmm, I wasn't thinking of using the flash as a writeback cache, I assumed you were talking about using it as a read cache. (Since overall, more reading is done than writing with most usage patterns.) The problem with using flash as a read cache from the hard drive is that you have to write something to the flash in order to read from it, and the flash memory's low write speed will be a huge bottleneck here.

Hmm, I think you'd pretty much have to use it as both a read and write cache if you want to achieve power savings, because otherwise even an occasional write will require you to spin up the hard drive really often (or leave it spinning most of the time).

As for the bottleneck, the I/O subsystem in the OS can mostly eliminate that bottleneck by reading stuff to RAM temporarily, then queueing it up to be written to flash later. A background thread could do the transfer from RAM to flash whenever the flash isn't being used for something else. If the background thread can't keep up with whatever is in the queue, just toss it -- it's only cached data anyway, which means by definition it's available on the hard disk. (And with I/O that heavy, you are probably going to have the hard disk running anyway.)

On a really busy server, this would probably not be practical since there might not be much spare bandwidth to the flash, but on a laptop, I would think it would work pretty well, because the flash should be idle much of the time.

Not that much of a slowdown

[Andy Dodd]

That disadvantage of FAT causes excessive seeking on mechanical HDs.

It's not a problem with flash. Flash seeks instantly and the amount of data that actually needs to be written to the FAT is miniscule. I'd estimate less than 5% overhead at worst, most likely much less for large files.

Simply put, the raw write speeds of flash memory are LOW. Flash is slow as hell when you want to write to it. They've made huge advances in read speeds, but not write speeds.

Re:Not that much of a slowdown

[AdriaanN]

Almost right!

Whenever you write a SMALL file, let's say 1 kb, you'll have to (re)write the block that contains the part of the FAT that holds the block number for that 1 kb as well.

In other words, depending on blocksize, when writing 1 kb, one may have to write double or even more (imagine a 4 kb block on a FAT32 system: 8 kb write for 1 kb).

The more the software driver buffers, the bigger the problem when the floppy or the flashcard is removed before everything is flushed. So, most drivers flush their in-core FAT whenever a file is closed.

What a POS

[Andy Dodd]

When accessing large files (not much seeking involved), modern hard drives have around double the sustained read/write speeds as that drive has CACHED read/write speeds. That thing's sustained write is only 1 MB/sec and its sustained read is only 2.3 MB/sec - even under incredibly heavy seeking, a decent SATA drive won't drop anywhere nearly that low.

Overrated

[Andy Dodd]

It provides minimal advantages (mainly battery backup) over simply increasing system RAM. For 90% of people it makes more sense to increase system RAM.

Now if that drive used an older memory technology that was much cheaper per megabyte than modern system RAM, it would kick ass. Unfortunately, you need to load it with the same stuff you put into your system, and unless you need the battery backup, you'll get much more performance by simply sticking the memory into your system itself.

Writing Cycles...

[gweihir]

... are the key. HDDs support practically unlimited overwrites. Today flash is somewhere between 10.000 and 100.000. Even with advanced defects management my guess would be they need at least 10 years and possibly much longer to be real competition to magnetic memory. Especially as long as computers use swap space.

This guy is just blowing smoke to make himself more important. Typical suit behaviour.

IDE interface

[marcosdumay]

Has anyone around here saw a flash drive that you can plug on a IDE (or SCSI) interface? That may be a very usefull step into replacing HDs.

Re:IDE interface

[Alderin1]

Actually, all you need is a CompactFlash device and one of these:
http://www.mini-box.com/s.nl/sc.8/category.14/it.A /id.243/.f

Pricewatch lists 4Gb Compact Flash cards already, and that top end is likely to rise quickly.

Re:IDE interface

[smoker2]

Search for it !

I want

[kadathseeker]

a solid-state device to run my OS and games off of, then a holographic drive to store everything on. Anyone like those new 150GB WD Raptors? I want one when they make it holographic, double the cache, and bump it up to 3MB/s SATA II. But most of all cheaper.

whats 30GB in 6 years?

[BibelBiber]

let's think back 6 years or so, notebook hds had less than 10 GB and compared to now this is pretty much nothing you would want. I mean this would not even be enough to have more than just the OS running on it. You want movies, music and pictures, all the space wasing stuff there is. In another six years you probably want 1TB of space on your disc to have all your HD-movies on it. That look a bit impossible to do (at least right now) with solid state storage solutions. Still I think the idea is good and solid state serves a great purpose for smaller devices such as iPods and the like.

Re:whats 30GB in 6 years?

[davidbofinger]

Hard drive capacity is doubling every, what, thirteen months? But the improvement in access speed is much slower.

How do the rates of improvement for solid state memory compare with this? Is solid state "catching up" to hard drives? Or just tagging along behind?

pt's rule #3

[mnemotronic]

The size of your current RAM was the size of your hard drive 6 years ago. The size of your hard drive now will be the size of your RAM in 6 years.

Re:pt's rule #3

[DeathFromSomewhere]

B.S. 512MB hard drives are early 90's, not late 90's.

Not so fast

[alienmole]

Flash memory isn't even remotely competitive with RAM memory in terms of performance, so flash won't replace RAM any time soon, not even in PDAs, never mind notebooks. It also has to be erased in blocks, i.e. you can't change a single byte at a time, which leads to serious inefficiencies if used as regular RAM. Finally, it supports a limited number of rewrites, so using it in place of RAM means that it would die really, really quickly. The reason they're only talking about using flash to replace hard disks in certain situations, like notebooks, is that because in other situations, it's just not viable.

However, a variation on your idea about "instant-on" is already being used by PDAs. Some PDAs back up their RAM to flash to achieve instant-on. There's a mention of that in the Wikipedia article on flash memory

"limited-write" memory

[peter303]

Most of these flash devices are limited to a hundred thousand writes or so. This fine for music or cameras, but not for routine computing. So you'll need add intelligence to the operating system to manage this limitation. And you'll still need unlimited-write memory for highly repeated computation.

Martians have been doing this

[peter303]

Recent NASA probes have been replacing old tape drives with solid-state memory to reduce the electric cost and unrelibiality of motors. The two Mars rovers use flash memory as well as the Saturn Cassini orbitor. Earlier probes like Jupter Galilleo and Voyager used tape drives.

Move along people, nothing to see here

[I Like Pudding]

Maybe you missed the RAM half of RamSan? It's a monsterous RAM disk backed with scsi drives, not solid state.

Go-L == Scam

[cwg_at_opc]

um, yeah. you don't work for them. riiiiiight. ;-)

Go-L is a known scam business, selling(?) over-hyped, under-powered equipment
as other-worldly, "killer" rigs. their web-site is a photoshop composite of product
images from other manufacturers/vendors with configurations that are at best, suspicious.
i was going to quote directly from their website, but it's really too much effort to refute
a scam-site when careful shoppers will see that everything is "pre-order".

here's a link from 2004:

http://www.notebookforums.com/showthread.php?t=142 10

and one from 2003:

http://thewolfweb.com/message_topic.aspx?topic=132 020

and from 2005:

http://plex.us/outbursts/liebermann.html

and if my poor memory serves me, there was also some noise about this on
some of the other enthusiast sites/forums; some review sites asked for
systems to review and benchmark but were 'rebuffed', to put it politely.
there was one actual review of a notebook they purported to sell, but overall, the
evidence is largely not positive.

their weird history of claiming ultra-high(read: impossible) performance systems
with conflicting equipment specs and configurations lead many to assume
they were fakers, or at the very least over-representing their products. there are
reports of buyers who paid large sums($20k) but never received anything.
i'd ignore these guys for now.

that said, there are solid state disks available for sale, but at extraordinarily
high cost. a google search turns up

Results 1 - 10 of about 1,790,000 for solid-state disk. (0.30 seconds)

which i won't even begin to quote here. ;-)