SSD... is anyone excited?

I've been loosely watching the solid state drives (flash based hard drives). And I was wondering is anyone else but me anticipating their mainstream/practical price points? they run on ata66 right now as far as i'm aware... if they put them on a sata 2 hookup and made em a little bigger I bet they'd almost kill the current HD market (save for mass storage/backup). Any thoughts on the actual transfer rates on such a setup?

I think they need to come down in price and improve their service life before I'd invest in one. Sure, they have their uses, but not enough to attract me yet.

Backing up the OS is the one main advantage I see in flash drives at the moment. Even with their limited lifespan, if you did a back up of the OS once a month, if by some disaster the original OS on a hard drive was lost, you could then reinstall the OS on a new hard drive with minimal trouble. Of course, that would mean having a flash drive of about 40 gig and keeping your data on a second hard drive for safety. But I keep my data on a second hard drive anyway for safe keeping.

I don't know that I'd do the automated backup. That would mean leaving it connected to the computer and I've lost a couple computers in the past due to lightening problems and storms. I'd do a manual backup and then disconnect the flash and put it somewhere safe from accidental electrical discharges.

1) They need to low their prices, with the same money I can buy 500gbs or 750gbs drives and get more space.

2) Raptors and Barracudas perform the same, sometimes better, and very few times slower but usually better. Why pay hundreds more for a couple of perfomance on only some applicationS?

3) They can be silent and everything, but that silence is not worth hundreds of bucks.

And that no will stay for a long time.... I know that

Yeah, I know. Well, maybe around $150.

unless i start making a rediculous amount of dough... I won't want one till they use a different connection.

Samsung is always first to the market with this kind of stuff... they're awesome. And as far as i know they manufacture everything in house, right down to the screws... so if they make a prototype, there's really no overhead save for the cost of making it and raw materials... I wouldn't be surprised if they even had operations to acquire the raw materials. :-p that's why they're so successful too, production costs are lower, and profits are lower, and they can put the same product on the shelf as other companies for a lower price and still make more profit... it's business genius really.

I would like to see drives setup where you read and write to the flash memory, and the flash memory reads and writes to a hard drive; sort of like a large buffer or cache.

I'm thinking of a setup like a SQL database where the fast read/write operations take place in transaction logs and write them to disk as they have time.

I think this is the way to use the SSD drives; I would buy one like this.

Another 2-3 years and I think they could start to become a much more popular choice. It's only that they aren't as cheap as the old spinning disk. That's obvious though, this is new technology. 10 years ago the largest hard drive was around 6gig.

SSD has jumped up the capacity ladder a lot faster than standard hard drives have.
Didn't some company recently announce a 128GB version? I swear it was only 6 months or so back that 32GB was announced as a breakthrough.
Who says they wont hit a Terabyte in 2 years?

I would definitely say that it is not the interface that is bottlenecking these meaning the ata66, if it was then the companies would simply switch it to sata2. I think it is the way the actual flash memory inside can read and write. Once they improve upon that, upgrading the interface would be nothing for them. Remember the technology for that is there and readily accessible. And to be honest, may even be cheaper to integrate that into the drives I would think just due to the mass quantities of it already produced for magnetic HDDs. Anyway, in answer to your question about speed, this is not a full blown scientific fact, but one could logically deduct that it is not the interface slowing things down, simply the read and write ability of the media itself. The one major advantage that could help speed this up is it's wide use. Meaning, flash drives, memory cards for cameras, all of which would benefit from speed boosts. Thus giving the companies even more reason to work on increasing the read/write speeds of the media.

I would have to think that the companies making these drives are going to try as hard as they can to keep the prices as high as they can, for as long as they can, and due to the limited number of companies that could make these in a quality manner, that could be a while. Plus, anyone buying these now are generally gonna be those people that want to get the latest and greatest, so I am guessing we will watch this technology step up slowly in terms of speed. That way the companies can max profit on the technology. I wonder if they are working on making the memory itself last longer. Given the fact of the finite number of read/writes that worrys me about using it as my main hard drive even if they do speed them up. At least with magnetic drives worse case scenario the data can usually be recovered if you want to spend the money. Does anyone know if that is the same case with flash memory or is it just that once the drive bites the dust so does your data?

I see your point, but if they stripe it, would you have the same issue as raid drives do? meaning that say one part of the drive crashes, do you lose everything? If not and they can just run it in a dual channel way then, yeah, that would be great to do. That could get the current SSD's to run at least sata 1.5 speeds, or close to it, definitely at least ata6 speeds.

StorageSearch.com has a lot of really good information on this subject. Anyone interested in SSD's should visit that site to get clear and accurate information.

Flash wear out is not a concern with SSD's. Someone at StorageSearch.com crunched the numbers and determined that a 160GB SSD drive with an 80MB/s throughput run at 100% write utilization would take 51 years exhaust the write limit of the flash memory. That's a brutal worst-case scenario. In real life, reads and writes average about 3:1, which means the predicted failure will take over 80 years.

Someone asked about mirroring flash cells to take up the slack in case one starts failing. Some manufacturers do include a certain percentage of extra cells that they write to when cells start approaching their write limit. Others have intelligence built into the devices that spreads the writes out across cells so no one cell gets hit more than another, effectively prolonging the life of the device.

The throughput issue is mostly about the speed of NAND memory. Currently the fastest NAND chips on the market run at 80MB/s, but the most widely used are closer to 55MB/s. This is why the ATA-5 interface is not a bottleneck. Samsung recently announced new PRAM with predicted speeds 30 times that of the fastest NAND, which is far faster than any interface can currently support.

For those of you with the budgets, take a look at Texas Memory Systems. They offer SSD SAN's and intermediate accelerator devices for high-volume and high-IOP (and high-end) environments.