Slashdot Mirror
Samsung's Solid-State Disk Drive Unveiled
Iddo Genuth writes "After unveiling their upcoming hybrid hard drive, Samsung — along with a number of other manufacturers — is planning to begin shipping solid-state drives during 2007. Unlike the upcoming hybrids, solid-state drives should work with windows XP as well as Vista." The drives will be introduced in 1.8- and 2.5-inch form factors for notebooks. While streaming performance can't equal that of hard disks, Samsung claims that random-access performance is more important and that (e.g.) Vista users would see a 4x speedup in many key operations. Pricing was not announced.
Now this is one configuration where this drive will make a large difference in bootup speeds. Office apps, audio, video and other media should be happy on the old 7200 rpm drives for a few years still.
Could someone tell me why one type of drive wouldn't work with a specific version of Windows? Shouldn't they be able to write drivers for that?
Adventures in Shaanxi
the days of the spinning platter hard drive are almost over. while it will take a long time for the SSD to completely displace all uses of traditional hard drives (especially in industrial storage -- where hard drives are now displacing tape), be prepared for an avelanche of new products in the 1-2 year time frame.
merlin
According to Microsoft, "SuperFetch understands which applications you use most, and preloads these applications into memory, so your system is more responsive".
Seems nice in theory, but the first thing I do to any XP machine that someone tells me is running very slow is to kill those quick start apps in the bottom right corner. Their use of processor and/or memory definitely slows the machine down overall. I'd much rather wait an extra second for an app to load so the system runs faster overall.
So they better have improved their techniques with this SuperFetch. If it causes many more context switches or reduces memory available to apps people are actually running then it'll be a hinderance. At the very least it should be automatically turned off for systems with less than an ideal amount of memory.
Developers: We can use your help.
Doesn't flash memory have a maximum lifetime (R/W cycles)? If so, are these new drives designed to "degrade" gracefully so that as the flash "rots", more and more data is stored to the drive instead of the memory? If so, this would mean that the drives would "slow down" over time right?
Reminds me of when a company in the 70's built a solid-state swapping "drum" memory system for IBM S/370 mainframes. Of course, that one wouldn't fit in a 2.5" form factor.
"It's the height of ridiculousness to say for those 9 lines you get hundreds of millions."
Coupled will fuel cell technology, mobile computing is finally going to live up to its potential.
And I love this William Gibson quote from 1991:
It wasn't until I could finally afford a computer of my own that I found out there's a drive mechanism inside- this little thing that spins around. I'd been expecting an exotic crystalline thing, a cyberspace deck or something, and what I got was a little piece of a Victorian engine that made noises like a scratchy old record player. That noise took away some of the mystique for me; it made computers less sexy. My ignorance had allowed me to romanticize them.The difference between stupidity and genius is that genius has its limits.
...because they don't want you to get a bad case of sticker shock. If texas memory systems (http://www.texmemsys.com/) is any guide, these things won't be comparable to platter drives in cost per GB per performance. Maybe they've figured out a way to manufacture the things not too expensively per GB but the performance will be wretched. And even though most apps will not care unless you have a stopwatch people will look at the raw numbers and shy away. Just see all the trouble AMD had with the Pentium 4 vs Athlon XP CPU GHz wars.
So now this might get Vista running half as fast as every other operating system, right?
This would be cool for your PVR solution. Faster, more quiet, uses less power, therefor cooler component, less fan noise, I hope. Now, can I strap one to my old Jornada 525?
Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong fix.
I did an eval of SSD back, oh five years ago for my employer. These were SSD's attached via SCSI to Sun boxes running Solaris and Sybase. Based on the results I saw then, I have two problems with this:
>Vista users would see a 4x speedup in many key operations.
Back in the day, we were seeing 10-20X improvements over spinning media in Random Access. 4x is almost not worth it, depending on price - give spinning media another year or two and they'll match that gain.
>Pricing was not announced.
Of course not, because it's going to be outrageously expensive!
-- "In order to have power, I must be taken seriously." -Mojo Jojo
I've got a fairly busy email server and this sounds like a great thing for the queue files... lots of little files, lots of random access.
Of course, the other posts about flash memory degrading after n writes would be something to watch, too.
The Army reading list
My notebook only has room for *one* drive onboard. I'm not going to replace a 80gb hardrive for a 4gb ssd (which currently cost $465 (see http://www.dvnation.com/nand-flash-ssd.html/). So the hybrid is the way to go ... but what I'd like to see is a
hybrid that just shows up as two drives under non-vista operating
systems. Then the boot stuff could go on the small flash drive and everything
else on the old fashioned (big) hard drive.
Wouldn't a better focus be on battery backed up RAM drives instead? Like those PCI DDR ram drives that cost a bundle. It would be nice to get a blazing fast PC3200 1GB RAM-Drive for $100.. which would be multiple times faster than these drives.
And if anyone had actually read the article, they would see that according to Samsung, the Flash technology in use in the drives has a lifetime of TEN years (your IDE / SATA HD likely wont last that long btw). They also note how much the R/W cycle issue has improved in the last few years.
/., we don't read the articles we just write silly comments first!
Oh wait, this is
"To strive, to seek, to find, and not to yield." - Tennyson
Flash is too restrictive in how many times it can be rewritten to be a viable replacement for a hard disk drive.
Nice try though...
I know far too little about flash to comment on whether or not it is adequate to replace a hard-drive, but I do wonder on a modern PC how many times does a hard-drive really need to be rewritten to?
Back in the olden days of computers (as in not that long ago) few people had enough RAM to keep an entire program in memory so the OS was constantly swapping data between Memory and the hard-drive. Being that we can now put 2 or 4 GB of memory in a PC (and in the near future 8-16GB of memory) how much swapping really needs to be done? If you use the sectors of flash reasonably evenly, how long would a flash hard-drive last? 2/5/10 years? How long is it reasonable to expect a hard-drive to last?
I understand they (Samsung) are the largest manufacturers of television sets of any kind now. And their stuff is of quality. Kudos to them.
It would seem to me that these drives if they were used might be present an issue with data security. Are there any plans to protect the solid state components from being read by unauthorized access? Hopefully the design is such that all data is protected but being new, I couldn't get enough details to make a determination.
Well, there's spam egg sausage and spam, that's not got much spam in it.
your IDE / SATA HD likely wont last that long btw
I've never had a drive not last at least 10 years. Are drives today made of lower quality?
Developers: We can use your help.
This might be perfect for page files if you have low amounts of ram and want to reduce the "Hard disk thrashing" that windows goes through when paging.
This is one of those interim solutions for early adopters who have more disposable income than capacity for delayed gratification.
Here's an "Ask Slashdot" moment though: why do the heads need to move at all? Why isn't WD or Samsung or Hitachi building a long, length-of-radius head over each platter? Then the only motor needed is for the platter, and the head is merely a fixed unit? This would probably reduce most HDD crashes too, since the arm would no longer traverse the drive plane.
I dunno, there's better ways to describe what I mean but I know there's something good in the creamy center of that idea.
-BA
Strange... the /. story doesn't say anything about it being flash ram. Not sure how you came to the conclusion that I didn't read the article.
Anyways, I know that R/W cycles have improved, but they still aren't at the point of lasting as long as hard drives, especially when portions of them are used for swap space, temporary files, and other virtual memory.
File under 'M' for 'Manic ranting'
Flash is too restrictive in how many times it can be rewritten to be a viable replacement for a hard disk drive. Nice try though.
Well, there's a nice unqualified anecdotal opinion, it must be proof. On youe average corporate desktop with word and excel and powerpoint and outlook and whatnot the laptop will die long before the HDD. Your average coder's debug files don't make a dent, not even the salesmen's multi-MB powerpoint presentations. Maybe, if the desktop was used for the typical P2P "download, watch, burn/delete" that'd keep writing huge multimedia files to the disk all the time, then maybe. Or if you construct some very extreme conditions like a 98% full disk with heavy swap, in which case I wouldn't trust a HDD's sectors too much either. As long as they're consistent with a predictable lifetime, I see no showstopper here.
Live today, because you never know what tomorrow brings
I weakest link will become the battery. Even after switching to OLED for displays and to solid state drives, the CPUs and the video cards will drain more and more power because they'll have to run behemoths like Vista on the machine. So unless there is a dramatic improvement in the basic processor design or battery technology (fuel cells?) mobile computing won't quite live up to it potential yet.
I've always wondered about this. Most modern flash seems to get 100k writes (many more reads). Fast flash is on the order of 13MB/s write.
With load balancing, you wouldn't notice a failure until all the locations were rewritten just shy of 100,000 times. So the drive will "fail" in once you've written 40GB of data 99,999 times, or almost 4PB of write ops. At 13MB/s, that's just under 10 years of 100% duty cycle writes. If you presume you'll read that data once at 20MB/s, and you allow only an 82% duty cycle overall (to make the math easy), then your drive should last 20 years.
I don't know about you, but I don't have any 20 year old computers or drives. The computer I had 20 years ago (PS/2 model 30, iirc) used 720k floppies, and a 20MB hard drive was a $400 option. Wait, check that. I do have a copy of Windows 1.04 on floppy disk here. It fits on three 720k floppies.
Is it just my observation, or are there way too many stupid people in the world?
This would make an ideal drive for streaming media servers and small databases, which is exactly what I currently need. Streaming media requires a lot of sustained reads from different locations, which taxes the ability of a drive head to cover. With 1ms access time, a single drive could replace a RAID configuration, saving power and space in our 1U boxes. Woot!
Whoops - I editied the post without re-reading it before I hit submit. I was assuming a 40GB drive. I also just read the second article and my numbers are off for their drives: let me update...
32GB, 32MB/s write speed, 57MB read speed. Assumed 100k cycles. 3.2PB at 32MB/s...8 years of service. I still don't have any 8 year old drives in my box-o-stuff, though.
Is it just my observation, or are there way too many stupid people in the world?
Since flash does have a limited number of writes, using one of these in a PC for daily use would be limited at best, so I'm wondering what types of applications would this media be ideal for?
The only answer that I could think of is anything that is 'write once, read many times'.
Movies - build a huge RAID array of flash drives an let them go to town on the lastest blockbuster.
TV - PPV system / VOD. New shows come on their own stack, plug them into the PPV system and be done with it.
Databases - certain tables that hold nonchanging data / lookup values.
That's about it. Anyone else?
III.IIVIVIXIIVIVIIIVVIIIIXVIIIXIIIIIIIIVIIIIVVIII
Let's look at the application, notebooks. There are quite a few pros for solid-state drives here: 1) HDDs are loud, 2) HDDs are hot (especially as you increase RPM), 3) HDDs are sensitive to motion, 4) HDDs require more power, 5) HDDs are marginally heavier (I mean the things are pretty small already). So the advantages here are pretty obvious, quieter, cooler, longer battery life, and marginally lighter notebooks.
Now, it is only fair we look at the downside, which is this overplayed write issue. Let us assume 10,000,000 writes (this is very generous, so I include 1,000,000 as well), since they will surely be using the best they can get, and this is pretty close to the high end that you will hear people discuss. You rarely re-write a vast majority of the software on your PC. Many programs are installed and never updated, and those that are updated are not done that often. If we assume, for the sake of sanity and argument, that the Windows system folder will only be written during Windows updates and that there would be one update per day that would be equal to something in the range of 27,000 days before you reached 10 million writes (only 2,700 if we say 1,000,000 writes). Most of your media files will be re-written even less.
So let us look at two things that can be written fairly often. First, you have a page file. The solution, load up your system with 1 to 2 GB of RAM and set the Windows page memory settings to the minimum. Of course, if Windows behaved properly, it wouldn't even write to the page file until AFTER the RAM was full (or damn near full). Second, user documents. Let's us assume your program performs auto-saves of your documents on a 5 minute cycle. So, the file is written one time every five minutes, 12 times an hour, 288 times a day (if you type for 24 hours of course), 105120 times a year (wow, I recommend some sleep and bathroom breaks), which ultimately results in 95 years (wow, congratulations on long life) of the file. Granted if we go with one million that is 9.5 years of continuous typing, but then you probably don't have much of a life if you are doing that.
This re-write claim is the most over-stated problem. Most places tell you the average life of today's HDD (for home use) is between 3 to 5 years, of course that is why they also tend to only warranty you for that long. Also, using my numbers, how many 9.5 year old drives are you using at home? Seriously, this problem is not that big of a deal; if it was going to be a huge problem, I am sure they would have though about that. (Note: This doesn't even get into the technology they use to spread the writes out to avoid wear.)
"Some days you just can't get rid of a bomb."
If you are going to use 1GB of RAM, it would be much more efficient to add this GB to your main memory and increase the size of your filesystem cache (if your OS doen't do it automatically). Ok, you wouldn't gain in boot time, but after that the OS makes sure your additionnal GB is used in the most efficient way.
It's almost the same thing as when pagefiles were introduced in Windows. People suggested using a RAMdisk to hold the pagefile, so that swapping would be much faster........
Hard drives last many many times longer than that.
File under 'M' for 'Manic ranting'
*IF* Windows behaved properly.
It doesn't. It won't. No amount of wishing will make it so.
File under 'M' for 'Manic ranting'
If you can't afford RAM you shouldn't even be dreaming about SSDs.
A hard drive can last well over 10 years... the only reason it typically gets replaced long before then is because of space issues, not unreliability issues.
File under 'M' for 'Manic ranting'
I've broken 32 gigs of applications ... over a year ago. I have a 100 gig windows/application partition. It is over half full. I don't believe I am the only one.
Amazing that this kind of stuff gets touted as innovative new features. I have in the past put together a shell script of a few lines which pre-loads commonly open files at boot time. It's trivial and shows just how inflexible Windows really is.
Deleted
But if both will last, statistically, through 5 years of continuous use, do you care that much that the hard drive will last to 25? Most people don't.
And the flash will fail non-catstrophically, unlike the likely outcome for your hard drive should it fail.
"Who is the Journal of Quantum Physics going to believe?" --Stephen Hawking
The transfer rate of most flash memory is still slower than hard drives. The advantage for now is no seek latency. However, transfer rates are improving pretty fast, so we'll see in that 1-2 year time frame.
If you look carefully at the photo of the HDD drive on top of a laptop in the article , you can read the popup on the screen saying '... to fix this problem'.
But if you write 32GB - 1 bytes data once and then rewrite that one byte 100000 the flash would fail in just couple hours (if we assume there are no reserved blocks that can be used for remapping bad blocks). Still, if take average of our values, i'd say 4 years is plenty for a laptop hdd. And ofcourse we have backups :)
- Raynet --> .
You can't install your OS to system RAM. Its also not easy to automatically have all documents or programs or whatever always loaded into RAM. Not to mention the limits on total physical ram (especially in the Windows world). If I could get an 8GB Ram drive for $400 I would probably do it. 3GB/s transfer rates blow any other currnet storage medium out of the water.
Now compare the price of a motherboard + Operating system + 8GB additional RAM(for windows users mostly) and you see how efficient adding more RAM is.
The usual way they construct it is like this:
1. Fill your drive 95%
2. Trash the remaining 5%. Your disk will now die in 1/20th of the time, that is a matter of months
IMO even that theoretical problem could be solved by active swapping, that is using some of your write cycles to move information internally. If you spent 100 of your 100k cycles doing that noone would notice. So when you're trying to trash those 5%, those 5% would swap places with the other 95%, even though there's no free space. For all I know maybe they do already, but if it was a problem that is the solution (this was sooo obvious. I bet it's patented).
Live today, because you never know what tomorrow brings
I call BS. You are still using all your 1gb hard drives from 1996? Or do you mean that they 'lasted ten years' as in, they lasted as long as I wanted them to last before I stopped using them and they are sitting on a shelf somewhere in my basement so I'm sure they work just great.
I just hope the drive offers a way to overwrite the flash part multiple times. I can see people making sure the hard disk is erased before moving a machine between departments or selling them, but then the next owner just checks the flash part, and finds document caches of documents which should never see the light of day.
Not true .. I only replace drives when they die. Although that could explain why my PC at home has 4 hard drives in it.
My System folder (OS X 10.4.8) is only 1.87GB and my Applications folder is 8.3GB (with iWork '06, Adium X, Textwrangler, One Button FTP, etc - aside from iWork, small programs).
Unless I'm missing some huge hidden folder, that means a 16GB drive would be plenty for most users as the OS+applications drive, unless (since I said "most users") Windows XP or Vista have become so bloated that they can't fit it all in even 16GB.
I've been wanting to build a computer for my ATV so I can have a nice mapping program (like TopoFusion, or maybe an open source app) and GPS tracking and recording.
These drives wouldn't be affected by the bouncing and vibration like a normal drive would.
Lose Weight and Feel Great with Isagenix
But they talk about Vista like it is already an existing product. (Comparing speeds, and stuff).
Patents Drive Free Software as Hurricanes Drive Construction Industry
It's a trap!
Whenever I've bought a new computer I've kept the older ones around for servers and "play". Right now my last desktop is a media server with a few old hard drives, together giving me a lot of space. The desktop prior to that is a web server. I never throw out old hardware that's still useful, including 10 year old hard drives.
Developers: We can use your help.
Being that we can now put 2 or 4 GB of memory in a PC (and in the near future 8-16GB of memory) how much swapping really needs to be done?
It depends -- Are you using emacs?
Backup early - backup often.
Too lazy to create a sig...
I've seen this problem happen in practice, with a development embedded system that booted off two flash cards. It ended up lasting 6 to 8 months or so before we needed to put in another flash card (it used CF cards, one 32mb which took all required R/W operations and one 1GB which was RO) and we took images for when the CF cards died. There was no swap, no journaling, and temp files were moved to a ramdisk where possible, but some files needed to be modified during development and the CF cards didn't last. Actually, some files were not readable after the flash failure. Perhaps this was due to an inconsistent filesystem state.
GPL: Free as in will
My 30GB IBM 75GXP is still running fine. I got that in around 2000. Not ten years, but then not too far off.
Do you have any better hostages?
So in theory it will go 10 years. In practice we'll probably have an industry wide recall by March 2007.
New hard drive technology is scary. Anyone remember the IBM Death Star? Yeah. Supposed to have newer faster tech. Replaced over 100 of them in our Dell Optiplex GX110's. I think I'll wait for the bleeding edgers to bleed a little before I even consider one of these.
I could hold my whole XP image on 10.4GB. Anything that's not on the image is on the corporate network somewhere such as Notes servers. So if someone would build a laptop with a 16GB fast SSD then that would be great. I'll even buy my own portable USB harddrive for everything that doesn't fit. For home use I already have a NAS.
It's at times like this that I wish there was a '+6 Funny'.
If these drives become standard they'll have a huge impact on my day-to-day.
The most common point of failure in a desktop PC is the drive head smacking into the disk platter in a rotational-magnetic drive. The worst part of these failures is that your drive head runs a real good chance of being over your important data when it hits (because you access it often, because it's important), so you're much more likely to toast your critical ACT! database instead of the rarely used Typing Tutor Turbo III you don't care about. Switching to solid state drives would increase reliability and reduce data recovery costs. You would see the impact on national GDP.
Many new algorithms, even today, have to be designed to minimize disk seeking and re-order disk accesses into sequential reads. Pulling a million 100 byte records from a database that are scattered across a 100GB table is enough to make an SQL database unusable since even if you're using an index and have it entirely resident in cache. Solid state storage would free up capital being sunk into this kind of development, which translates into increased investment in more application features or lower costs.
Once the streaming performance is addressed (matter of time), solid state drives are going to be the standard and we'll wonder why we ever put up with rotational-magnetic storage.
There will be plenty of swapping, as program designers will bare witness to the multitudes of available RAM by creating ever bloating application code and useless background services! At least with solid state there will not be much weeping and thrashing of disks!
Do you really have PC systems or servers that have 95% of the HD space as static storage? I usually make sure all of my machines are no more than 75% full, and preferrably 50% or less. On traditional HDs, small space means lots of fragmentation and slow machines.
Besides, with a simple utility you could "defrag" your flash drive occasionally and have the internal management software/firmware do a remap to get files which are over x days old into memory which has a disproportionate number of cycles. Further, laptop users are used to part wearing out over time. Anyone who runs on battery frequently knows that after a few hundred cycles the batteries are shot. I've taken thinking of each battery-usage cycle as a $0.25 charge for being cordless, and I've already been through one battery on my machine that's barely 2 years old, and my second is probably down to 80% max cap already.
Is it just my observation, or are there way too many stupid people in the world?
We still have a few SPARCStations and PoewrMacs that are more than 10 years old, but I couldn't tell you the ratio of working to dead disks, since the ones with broken hard disks were thrown in a skip before I got the working others. I'd say 10 years is more than enough for a hard disk drive, especially if all you lose is write-ability at the end of that. After ten years, you can easily buy a much bigger drive and copy all the data on to that. I'd probably want to every five years anyway...
I am TheRaven on Soylent News
"Are drives today made of lower quality?"
According to a buddy of mine who works in data recovery, yes. It's also not so much a question of quality as much as it is a question of platter density. According to him, a single platter 20Gb drive lasted a good long time. In the data recovery business, they're seeing a lot more high density drives than the older ones (even though you'd think the older drives would fail more frequently). (why do you think the HD companies dropped the warranty down to 1 year in the last few years?)
Personally, I must have bought about 30 HD's in the last 10 years, ranging from 4Gb in '97 to four 300Gb SATA drives two months ago.
I guess I've had at least a third fail in one way or another. I'm not talking about dropping them or anything, they just quit working.
How's your track record? How many drives have you actually owned, and fully used, and how many have failed? I'm curious.
"If you could only see what I've seen with your eyes..." - Roy Batty
Seems to me that you could do RAM+flash; have it work as a RAM drive when "powered on", but then when powered off (either with the whole system, or by power management powering the drive down due to inactivity) it dumps the RAM to the flash, and restores the RAM from flash when powering up. You get better performance, and save read/write cycles on the flash (of course, it'll be much more expensive than a flash drive, too.)
You might ask "why not just get more system RAM", and of course, that's a viable approach. OTOH, this way you might save money for the amount of fairly-fast storage by getting RAM that's not as fast as you'd want for system RAM, but still faster than reading from a traditional harddrive or flash. Of course, given the size of RAM modules, it won't be good for the "main" drive except in specialized applications, but it might be useful for special-purpose drives where access speed is critical.
4x is almost not worth it, depending on price - give spinning media another year or two and they'll match that gain.
The thing that increases for HDs is throughput. Essentially this means the sustained rate that can be transferred by a HD once it's found the right position to read from.
What DOESN'T get better very fast is latency. That is the time it takes for a HD to seek a new position on the HD.
So a 4 times improvement in load times is extremely significant, and worth the money. I'd love to have about 5-10 gigs of space for the OS to load up very quickly (or to make things faster, from a hibernate/deep sleep). I'd probbably buy such a HD and leave any major data storage to my fileserver.
It remains to be seen how much this actually will improve performance however.
Of course not, because it's going to be outrageously expensive!
I bet it won't. This is really just taking two pieces of off the shelf technology and combining them together. I'd expect a 20 gig drive to be somewhere around $300-$400. That's certainly a lot more expensive than a HD, but it's not so expensive that it doesn't fit into the budget of higher-end computers.
AccountKiller
I can't recall exactly how many I've used. I've owned at least 10 drives in the last 20 years. I remember only one that had bad sectors spreading quickly after many years of use. Most of the rest were used regularly for 8 or 10 years.
Developers: We can use your help.
Nah, the internal memory-management load balances the writing across the whole flash memory range. In your example you would just write to 100000 different bytes. Also, a journalling file system is different as well.
did you remember to mount everything read only and/or use "noatime".. otherwise there's a "hidden" write whenever you access a file or directory
at least you didn't use swap on a flash drive, I don't think the wear levelling would cope too well with that; I wonder whether the flash controllers understand the use of FAT and optimise accordingly - people with Zauruses who put ext2 on flash seem to have quite a few problems.
Yup, this project spanned multiple years and quite a bit of effort went into ensuring a good lifespan for the cards. They weren't exactly brand name cards though, we just bought a lot of cheaper ones. The device was under development, so many things ended up changing from release to release. Some storage was also needed for data acquisition, which was certainly part of the issue.
GPL: Free as in will
most people turn off their POST memory testing
Are you really claiming that most people go into the BIOS and change the defaults to save a second off of boot time? I find it very hard to believe that any significant number of people change that setting even among people who know where that setting is. System bootup generally takes at least 30 seconds on any system. The memory test isn't going to take more than a second or two unless you put a ton of ram into a really old computer.
Nah, the internal memory-management load balances the writing across the whole flash memory range. In your example you would just write to 100000 different bytes. Also, a journalling file system is different as well.
Load balancing writes only works if the flash has a significant amount of unused space. In the grandparent's example, if you fill the flash and then change the same byte 100,000 times, you can't do any load balancing. There just isn't anything to spread the load over.
Of course, the answer to that is to have hidden reserved space. Ok, that works, but how much reserved space do you need? 10 additional bytes just means you need to repeatedly change 1 bytes instead of 1 to burn out the flash.
[Change bytes to sectors/blocks/addressable units or whatever you see fit in the above. The units are irrelevant.]
Anyway, I'm not saying load balancing doesn't help things greatly. Just saying that things aren't as simple as you're making them out to be.
Of course, if Windows behaved properly, it wouldn't even write to the page file until AFTER the RAM was full (or damn near full).
Why do you believe this is "proper" behaviour ?
But screw Flash. PRAM (Or OUM, call it what you want) Its going to be awesome. You can use it for nearly everything. l1/l2 caches, solid-state drives with higher data densities because of the ungodly small form factor possible with teh technology, system memory. Hell, you could just as easily build a computer with a real "plug and play" OS (install the OS on a stick of RAM, plug it in, and turn on the computer. Computer accesses the OS RAM bank, near-instant OS load (or at least significantly increased loadup times, by far.) I say PRAM is the way to go. Trillions of write cycles? Bye, flash memory.
Still waiting on Serviscope_minor to wake up to fucking reality and realize that Jessica Price isn't going to fuck him.
Humm, if we assume that the flash size is 32GB and I write there 32GB data, then how can it load balance anything? Except using some reserved blocks that I cannot use directly?
- Raynet --> .
There surely has to be a better way than that...
I mean, if such an archaic and painful method is the solution to defragging for linux... What's the point?