Sign in to follow this  
uart

Are SSD's poised to take over in terms performance?

Recommended Posts

We all know that solid state drives have come from miles behind mechanical drives in STR to be just about level pegging right now. I can't really see any reason why SSD speed won’t just continue to rise, soon surpassing all desktop drives and not too much later even the best enterprise and 15k scsi drives.

Do you think this will soon happen? Once the STR surpasses the best mechanical then SSD will be the undisputed performance leader in all categories due to the low access times.

I don't see SSD's being comparable to mechanical storage any time soon in terms of cost. The cost per GB of SSD's is falling rapidly but so is that of rotational storage. But none of this will matter once the SSD is undisputed performance leader in all categories. There are a good many power uses that will have no qualms dropping hundreds of dollars extra for the performance god.

I’m thinking as little as one to two years and we might see pricey but undisputed “Raptor Killers†targeted squarely at the performance market start to pop up. Do you agree that SSD's might very soon be making serious inroads into this market segment?

Edited by uart

Share this post


Link to post
Share on other sites

We all know that solid state drives have come from miles behind mechanical drives in STR to be just about level pegging right now. I can't really see any reason why SSD speed won't just continue to rise, soon surpassing all desktop drives and not too much later even the best enterprise and 15k scsi drives.

Do you think this will soon happen? Once the STR surpasses the best mechanical then SSD will be the undisputed performance leader in all categories due to the low access times.

I don't see SSD's being comparable to mechanical storage any time soon in terms of cost. The cost per GB of SSD's is falling rapidly but so is that of rotational storage. But none of this will matter once the SSD is undisputed performance leader in all categories. There are a good many power uses that will have no qualms dropping hundreds of dollars extra for the performance god.

I'm thinking as little as one to two years and we might see pricey but undisputed "Raptor Killers" targeted squarely at the performance market start to pop up. Do you agree that SSD's might very soon be making serious inroads into this market segment?

Share this post


Link to post
Share on other sites
We all know that solid state drives have come from miles behind mechanical drives in STR to be just about level pegging right now. I can't really see any reason why SSD speed won't just continue to rise, soon surpassing all desktop drives and not too much later even the best enterprise and 15k scsi drives.

Do you think this will soon happen? Once the STR surpasses the best mechanical then SSD will be the undisputed performance leader in all categories due to the low access times.

I don't see SSD's being comparable to mechanical storage any time soon in terms of cost. The cost per GB of SSD's is falling rapidly but so is that of rotational storage. But none of this will matter once the SSD is undisputed performance leader in all categories. There are a good many power uses that will have no qualms dropping hundreds of dollars extra for the performance god.

I'm thinking as little as one to two years and we might see pricey but undisputed "Raptor Killers" targeted squarely at the performance market start to pop up. Do you agree that SSD's might very soon be making serious inroads into this market segment?

You must be referring to NAND based SSD's?

DRAM based SSD's (like Texas Memory Systems, Hyperdrive4) already are undisputed performance leaders both in terms of STR, low access times, and $$ :)

Current NAND SSD's with in almost reasonable price ranges have poor write speeds, with the exception of MTRON's (but is priced significantly higher)

Samsung 2.5" SSD $999 58MB/s Read, 32MB/s Write

MTRON 2.5" SSD $2500 100MB/s Read, 80MB/s Write

Share this post


Link to post
Share on other sites
You must be referring to NAND based SSD's?

DRAM based SSD's (like Texas Memory Systems, Hyperdrive4) already are undisputed performance leaders both in terms of STR, low access times, and $$ :)

No I'm not referring to DRAM drives, I think only non-volatile storage can be a real replacement for current HDD's.

Current NAND SSD's with in almost reasonable price ranges have poor write speeds, with the exception of MTRON's (but is priced significantly higher)

Samsung 2.5" SSD $999 58MB/s Read, 32MB/s Write

MTRON 2.5" SSD $2500 100MB/s Read, 80MB/s Write

Yep but look how far they've come. A few years ago flash was way slower in STR and now it is biting at the heals of HDD's. I know write is lagging a bit but read is more important for most application anyway. I think in about one years time read STR will surpass desktop level HDD, writes will be comparable and prices will be about 1/2 of current. Add one more year and the performance market for flash SSD's should be huge.

Share this post


Link to post
Share on other sites
You must be referring to NAND based SSD's?

DRAM based SSD's (like Texas Memory Systems, Hyperdrive4) already are undisputed performance leaders both in terms of STR, low access times, and $$ :)

No I'm not referring to DRAM drives, I think only non-volatile storage can be a real replacement for current HDD's.

Agreed. DRAM SSDs also can't compete on raw capacity. You can cram 128GB of flash in the same space as one GB of DRAM.

Current NAND SSD's with in almost reasonable price ranges have poor write speeds, with the exception of MTRON's (but is priced significantly higher)

Samsung 2.5" SSD $999 58MB/s Read, 32MB/s Write

MTRON 2.5" SSD $2500 100MB/s Read, 80MB/s Write

Yep but look how far they've come. A few years ago flash was way slower in STR and now it is biting at the heals of HDD's. I know write is lagging a bit but read is more important for most application anyway. I think in about one years time read STR will surpass desktop level HDD, writes will be comparable and prices will be about 1/2 of current. Add one more year and the performance market for flash SSD's should be huge.

Also that Samsung price is heavily marked up; you can find them for sale at the $480 mark already.

And yes, read speed is more important because it can't improve as effectively with caching. Whereas a small amount of DRAM will make write speeds more than fast enough for real world use.

Share this post


Link to post
Share on other sites

not for another year atleast the disk makers don't want to instantly go out of production!

it also takes years of setting of standards & testing to get to the level of reliability we have at the enterprise level that is going to be a while before alot of the current SSD attempts (the cheaper ones) are accepted.

at the end of the day SSD won't become mainstream until they are 64gb+ and $200...till then they can just keep on playing around

Share this post


Link to post
Share on other sites
... at the end of the day SSD won't become mainstream until they are 64gb+ and $200...till then they can just keep on playing around

Well considering that even 74GB Raptors are still like $150 plus, if an SSD was available with substantially better performance I'd say it would command $200.00 very easily. That's kind of the point I was making, if SSD's double in speed and half in price then they're going to be irresistible to the performance market. How long will that take to happen? Given there recent rate of progress it shouldn't be more than 2 years.

Share this post


Link to post
Share on other sites

It's nice to think that NAND Flash-based SSDs will be able to outperform HDDs in performance as well as price. However, I really don't believe that this will occur, due to limitations of the technology.

I believe we need a breakthrough technology to really boost capacity, performance, and lower the price point of SSD to a level where they can really be compared to a normal HDD.

That's just my opinion, though.

Share this post


Link to post
Share on other sites
It's nice to think that NAND Flash-based SSDs will be able to outperform HDDs in performance as well as price. However, I really don't believe that this will occur, due to limitations of the technology.

I believe we need a breakthrough technology to really boost capacity, performance, and lower the price point of SSD to a level where they can really be compared to a normal HDD.

That's just my opinion, though.

Yeah I know that SSD is not going to compete on a capacity/cost basis in the foreseeable future. But what about performance? Do you see any reason why SSD performance won't continue to grow and thoroughly surpass conventional HDD's in the next few years? The point I was trying to make is that if this happens they can still have a big market even if they’re not competitive on GB/dollar.

The real question here is does anyone know of any reason why flash based SSD's performance wont continue to climb and soon outpace conventional HDD’s.

Share this post


Link to post
Share on other sites

Most money to be made is:

- SCSI/SAS drives for OEM servers

- IDE/SATA drives for OEM desktop machines

Until the major OEMs can get reliable and reasonably priced access to such drives, they'll be niche products at best.

Reliable access means an uninterrupted stream of units.

Reasonably priced means "as cheap as possible" for a desktop and "not too damn expensive" for servers.

These two things will be more important than technical superiority (Betamax anyone?). And if the major OEMs don't buy shitloads of such drives, the price won't come down a lot because there will be little economics from scale. It's that chicken and egg thing. Same as IDE and SATA. SATA only got really mainstream when major OEMs (Dell, HPaq) started including them in their machines and that was only a different physical connector. The drive media itself didn't change one bit - pun intended.

Share this post


Link to post
Share on other sites
Most money to be made is:

- SCSI/SAS drives for OEM servers

- IDE/SATA drives for OEM desktop machines

Until the major OEMs can get reliable and reasonably priced access to such drives, they'll be niche products at best.

Reliable access means an uninterrupted stream of units.

Reasonably priced means "as cheap as possible" for a desktop and "not too damn expensive" for servers.

These two things will be more important than technical superiority (Betamax anyone?). And if the major OEMs don't buy shitloads of such drives, the price won't come down a lot because there will be little economics from scale. It's that chicken and egg thing. Same as IDE and SATA. SATA only got really mainstream when major OEMs (Dell, HPaq) started including them in their machines and that was only a different physical connector. The drive media itself didn't change one bit - pun intended.

I don't completely agree with that. Otherwise, how do you explain the success of the Raptor? Major OEMs haven't spec'd the Raptor in high enough volume to make an impact. You may call the Raptor a niche market product, but it's a high margin niche product. An SSD in a similar price and capacity range as a Raptor, with better performance would be a hot item for computer enthusiasts everywhere. The question goes back to the original poster's thread topic... How long until that happens?

I for one would love SSD's for both of my home PC's. I have a Raptor in my main PC for it's high performance. I would love to replace that with an SSD with better performance, and cooler and silent operation. For my HTPC an SSD would be perfect, again for it's cooler and silent operation. In both cases I have network storage for most data storage, so the SSD doesn't need to be huge.

Another area I could see the SSD having huge success is with corporate laptops. The two major points of failure with laptops are the hard drive and LCD. Eliminate the HD as a major failure point, while dramatically increasing performance and power efficiency and I expect sales of SSD's to OEM's to be quite substantial. Also, laptops are becoming much more common with home users and even gamers. Get rid of the major bottleneck in the laptop (the HD) and a laptop becomes much more appealing to many people.

Share this post


Link to post
Share on other sites

In addition to all the benefits mentioned in this thread, it's also interesting that SSD drives in laptops should give the same level of performance as SSD drives in desktops (whereas laptop hard drives are traditionally slower than desktop hard drives).

That got me to thinking though, what will be the most popular form factor for SSD drives? The 3.5" SATA/PATA, or the 2.5" SATA/PATA (or 1.8", or ExpressCard)? Or a completely new size and interface?

Share this post


Link to post
Share on other sites

bennt

The volume of Raptors is rather small compared to 7200 rpm IDE and SATA disks and it uses proven technology. It's also the only 10k SATA drive which very likely contributes to the fact that major OEMs don't sell it (only Dell does I think) because they'd be tied to a single supplier.

I agree about laptop drives though. Besides speed, an SSD would be less vulnerable to shocks, even in use.

Share this post


Link to post
Share on other sites

I don't see any point in introducing a 3.5" SSD - you can easily fit a 2.5" SATA drive in most desktop PCs, and for the rest you can use an adaptor. We're already seeing a mixture of 1.8" and 2.5" SSDs being released, but as long as 2.5" allows the largest capacity and the lowest price (less miniaturisation), it'll lead in the desktop segment. 1.8" drives will be popular in laptops, but are harder to integrate into desktop systems at present. Ultimately, I expect the industry to transition to 1.8" drives , starting with laptops, then servers, and finally desktops. But desktops haven't even transitioned to 2.5" drives yet, so that's still a few years away.

The exciting thing about SSDs is that it's so much easier to perform a form of internal striping to extract twice the STR by using twice as many chips. There's plenty of space for that. So it shouldn't be hard for manufacturers to come up with faster drives, even if the performance of individual chips doesn't increase much.

Share this post


Link to post
Share on other sites

I have been waiting so long for these. I think they will really take off once people see a noticeable speed boost. The silent and more reliable and more shockproof are no-brainers. But I think speed will really sell people.

At the current price of about 500 for a 64GB instead of 32GB, I would probably spring for one. Of course, if that were 250 or even $300, well, that would be a no brainer for me.

Imagine being being able to format in a second, and never needing to defrag, but if you did it would be a matter of seconds. Imagine a computer that is really quieter. I have been a huge fan of quiet computing for years. So, this SSD drive would make a very noticeable difference in noise, speed, heat, reliability (no moving parts, how cool is that?)

I hope SanDisk and Samsung get on with production, because unless you are Dell, these things are scarce as hens teeth, and how can the price go down if no one can find them to buy them in the first place.

I have tried to call both SanDisk and Samsung to try to buy say, 24 of them, and neither one can help me at all. The only place they are for sale is Ebay, and I would kill to know where they are egtting them :)

I think these will catch on really really quickly. 32GB is good for most laptop users I know, and 64MB would be good for the rest, and if you are a heavy big file user, just off load to a portable USB drive.

Folks, I think any one who has thought that price will keep people from buying these will find out the very definition of WRONG super quick. Remember, this is the culture where people are LINED UP ALL OVER THE COUNTRY TO BUY A $500 CELL PHONE THEY HAVE NEVER EVEN SEEN IN REAL LIFE YET.

I have been wrong before,, after all, I really thought people would flock to VISTA, but I believe in my bones I am so right about the adoption of these SSD drives.

I used to sell simple RAM at Egghead Software for FIFTY DOLLARS per MEGABYTE. With no price break even for FOUR MEGABYTES. Have a look at the price of RAM today compared with that.

Thanks for reading, Dave

Share this post


Link to post
Share on other sites
I used to sell simple RAM at Egghead Software for FIFTY DOLLARS per MEGABYTE. With no price break even for FOUR MEGABYTES. Have a look at the price of RAM today compared with that.

Hey I'm old enough to remember those bad old days. It's just amazing to look back at the crazy prices we used to pay isn't it. I remember my brother buying a 25MHz 486SX with 32MB of ram and it cost him about 4k (probably more like about 10k in today’s money). About half the price was memory alone, 32k was considered "huge" back then. To make it even worse he then had to go drop another $800.00 for a maths co-processor just to get a lowly 10MFlop FPU (back then most "desktop" systems didn't even come with any FPU at all, it was considered something only needed by "professionals" and a hefty cost upgrade).

It was just at about that point that the RAM cartels were getting broken and AMD and others were entering the market with "clone" 486's because about 6 months later I bought a 66MHz 486DLX (AMD 486 clone with FPU included) and a motherboard and 64MB of ram, all for about $700.00. I can remember my brother nearly falling over in disbelief when I told him what I payed, he really upgraded at a bad time.

Edited by uart

Share this post


Link to post
Share on other sites

Last month MOSAID announced a new controller/interconnect technology that's capable of delivering up to 800MB/sec sustained to an array of flash chips. Pretty cool.

http://www.mosaid.com/corporate/news-event...2007/070507.php

They call it HyperLink NAND, built on a high speed serial point-to-point daisychain layout. Makes sense, philosophically similar to HyperTransport and FBDIMMs really. Plus they also claim to have new program/erase algorithms to make writing as fast as reading. Interesting stuff.

Share this post


Link to post
Share on other sites
Last month MOSAID announced a new controller/interconnect technology that's capable of delivering up to 800MB/sec sustained to an array of flash chips. Pretty cool.

http://www.mosaid.com/corporate/news-event...2007/070507.php

They call it HyperLink NAND, built on a high speed serial point-to-point daisychain layout. Makes sense, philosophically similar to HyperTransport and FBDIMMs really. Plus they also claim to have new program/erase algorithms to make writing as fast as reading. Interesting stuff.

Buy that stock.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this