There are probably some real damages, in terms of distress and so forth, with hard-to-quantify monetary value. However, if the scam was deliberate, punitive damages are certainly called for.
Re your "correction": there are plenty of/consumer/ buyers that are insensitive to price below a certain price point, and for some of them, that price point is well over $50 when it comes to things like consumer electronics. I'd say my own personal wallet would open up for a card of that capability, were it available today, in the $300 range. Honestly.
I see what you're saying about sensitivity to data locality. While there is unfortunately as of yet a solution for this, what's wanting here is "transparent storage migration" (on a block level). Think logical volume that is data access frequency aware and likewise aware of drive characteristics. I.e., just make sure the blocks are where you need them most. Not a hard algorithm to write, just needs to be a market need.
While I am indeed sure that the "wanting" thing above will see solutions, I could see how they are unlikely to see the consumer light of day. I'd be more likely to expecting things like hybrid drives featuring something like fusionio technology internally, for a mega cache of sorts. At the right price point, of course. This leverages all sorts of sensible economies of scale, like being able to plug the drive into your favorite RAID controller.
Be that as it may, consider. Is there not some near feature, not so far off day where software itself would be unlikely to be large enough to fill up a drive like this, for many/most users? I think so. Also, there are many classes of work where the job being worked on does fit reasonably well on the drive, and would benefit. Digital image production, various kinds of (small) streaming IO problems and so on.
Anyway, I agree that the probable use of these cards are for cache. I'm just saying that there are plenty of applications that would benefit, should the price point be low enough, and there would certainly be buyers,/regardless/ of the ratio of price to bulk storage.
In enterprise space, this is obvious from the bevy of buyers who eat up 15K FC/SCSI/SAS drives just to scrape out every last ounce of IOP. Those drives are/capacity limited/, man.
I consumer space, it's also obvious. Just look at the enthusiast market.
There's also the workstation space; people with a capital budget, but are sensitive to labor hours.
As far as whether or not their performance claims are lies or not, I can't say. At the moment they are not available, and I'm not at the point of wanting to test prerelease silicon. What I will say is that if they promise "sustained read rates of 800mb/s" and don't deliver in good faith approximately that, I can arrange some very bad mojo for them.:-)
I do feel sure of is that it's possible to do what they say. They really just need to do simple striping of the nonvolatile RAM devices on that card. The way they are doing it has its limits (its scale-limited), but it's also economical. While they could make a 300mB/s SAS form factor, all buyers would then have to buy the RAID card in addition to their drives. Instead, they've married the "RAID" and "drive" circuitry on board. Interesting choice. Says a lot about their intended market.
I think that market is this: enthusiast, and add ons to enterprise storage products (EMC, NetApp, etc). They might seem some workstation buyers also.
And if I had a choice between say, one of the new 64GB SATA flash drives for $50 USD, or a 500GB 7200rpm Seagate SATA HDD for free, I'd go for the Seagate.
There are many buyers who are not like you. The issue is that to many buyers, both $50 and $0 are "free". There is a price point threshold below which the cost is a non expense. This is particularly true in enterprise purchasing situations, where the processing of the paperwork to merely buy and item is hundreds of dollars. While that only assesses the impact of one whole purchase, these and other factors are very real. They are also real in the consumer space, where many classes of buyers routinely behave exactly the same way. I would happily pay $50 for a device that is 8X faster than a hard drive for read and 1000X faster for random access. I already have a 500G hard drive, and it's not full...
I did not imply that you said RAID-0 would solve everything. Rather I am saying that it can be quite difficult to nurse high read/write rates out of RAID structures that don't lose large amounts of data. RAID-5 and RAID-6 won't match the read/write rate of one of fusionio's cards in any configuration from any enterprise seller of storage.
You're right, by the way. I am accepting their read/write rates at face value. This shouldn't trouble you, its a discussion about a theoretical, not-on-the-market product. Let's just suppose.
It might surprise you to learn that we are considering using their product in a clustered storage solution at their listed price. A couple grand per card does not seem out of line to me, as the backing device for a persistent cache. If the product performs as advertised, of course.
I poked around looking for RAID-2 information. Both striping and some sort of error correcting code are mentioned, but otherwise this solution is not discussed much. Tell me more. What are the read/write performance figures for a group of 10-14 drives in this configuration? And what happens if you lose a whole single drive?
14 15K FC won't give you this level of performance, unless you go into RAID-0. I'll discount RAID-0, because it's almost never used in real deployments.
I feel certain that this class of device will appear, and quite soon, in enterprise storage solutions where it will be used as a persistent backing store (cache) in the very RAID arrays that you are talking about. This isn't just guesswork; my position in industry is such that enterprise storage vendors do backflips in order to show me their developing products and roadmaps.
You're losing site of the forest for the trees if you think that the ratio of the costs is the deciding factor for purchase. It's absolutely not. Think it through. If a 64GB device like that were $50, would it bother you to pay for one even if you get (one) 500GB drive for free? The ratio is infinite, but I doubt you'd be bothered, and feel sure that you'd think long and hard about forking over the $50.
You'd think longer and harder if you knew just how hard it is to get reliable, sustained, real-world read rates of 800 MB/s...
Doesn't have the ram, but then given its performance figures you shouldn't care (and if you do, let's not forget you're asking it to do what your OS already does). Same goes for your write-back: at 600MB/s, why?
For example, this device here http://www.fusionio.com/ if affordable enough, would beat any RAID array you're ever likely to put in a workstation. In fact, it's hard to beat a device like this with ANY commercial RAID array. NetApp 3070's can't pull those numbers, and when they manage to get close, you're talking 14 15K FC drives.
Other "features" of your RAID array might be questionable as well. Like the power. And the noise.
One final factor. At SOME price point the fact that there is a ratio difference in price points becomes irrelevant. If you could get a 64GB fusionio card for 20 bucks, you wouldn't really care that a 64GB hard drive is 20 cents. The 100:1 difference in price is nothing, because both prices are nothing. At that scale and price point.
"Consumer price insensitivity" is the term. At some price, the price doesn't matter.
I'd be interested to hear of people who are running Oracle into things like Lustre. It should work, and work well, but I suspect it will take a brave man to sign it off into production.
The problem with Lustre is that if you configure in the fashion that one usually uses Lustre, there is a small non-zero chance that something can happen to your cluster such that you can realize an availability problem that may impact the entire tier. A future version that does the distributed stripe thing will address that.
(Note that there are still good uses cases for Lustre regardless of the possible tier availability issue, and for the most part the problem can be addressed independently by having each Lustre node be itself HA; note however that in all standard HA patterns, total loss of the storage chassis still spells doom to recent non mirrored data... BTW, if you think storage chassis loss isn't necessary to account in your availability model, then you're golden... depends on what you're doing).
If you're interested in clustered storage ala Acopia, but need it on a SAN, talk to the people at XIV Storage. They're doing clustered storage for block-level data, and are able to easily keep ~40gbit/sec fed. Course, this is a startup. Magna caveat emptor.
All good thoughts. Agree and all that. This being slashdot, instead of talking at length about agreement we'll have to up the ante and disagree about something (hah). Let's talk about that "you still need to mirror the SSD until your insane" comment. I'm finding in-box redundancy to be less and less useful. Check out http://www.isilon.com./ These guys, and emerging vendors like them, are beginning to move away from in-array replication of data to cross-array/cross-box replication. Packets are striped across the filers, instead of inside of the array. The beauty here, is that you can actually lose the filer chassis entirely, and you don't lose any data. You can replace when convenient, and get closer to the idea of "ignoring the chance of RAID-5/RAID-6 simultaneous disk failure". It's still not entirely there, mind, but since RAID rebuilds involve any disk in all of your racks, you have no need to do anything, the system is essentially self-healing.
Soon, Lustre file system will have such a feature. I can't wait.
I'm sure you will find more and more turnkey cluster storage vendors hitting the market soon, precisely because clustered storage solves bandwidth, capacity, and manageability problems so well. All sorts of serviceability problems are cleanly solved by breaking away from classic RAID and moving to the right kind of cluster architecture.... problem is, until recently it's been a science project. Turnkey cluster appliances are right around the corner (or here, if you can live within Isilon's NAS-only constraints).
It's not all one way or the other. At some price point, there are consumers who will buy the fast thing, just to have SOME of the fast thing. One is not required to have only one thing inside one's computer. For example, I could make productive use of a 32 or 64GB flash system if it were fast enough and cheap enough RIGHT NOW, regardless of the fact that I have a 500GB SATA drive in my system in order to store movies and images.
There are also valuable business applications for the same technology. If 64GB flash were "relatively affordable" and noticeably faster and more effective than a good RAID array, then these flash drives could be an important component of an enterprise storage system. Think "cache times ten, plus non volatile."
I'd love to build a Lustre Cluster where each of the nodes had one of these as a persistent backing store for cache: http://www.fusionio.com./ As an added plus, their stated market price of $30/GB is ALMOST worth it.
An investment of equal size has a return of it own. That return COMPOUNDS. A mutual fund will average about 11.5% annually, COMPOUNDED, when you pick a simple S&P 500 index.
You have NOT performed the calculations. You've only performed the half of them. You calculated, on the back of a napkin, so to speak, how much money you would "save." You did not calculate, however, how much money you would fail to gain by not investing that money in good investments elsewhere. Not to mention: with the "good investments elsewhere," your principal is still in your hands at the end. With your current strategy, the money is not.
A quick shot in excel shows me this. $35,000 at 11.5% interest is worth $100K after 10 years. Did your 10 year calculation show a similar return on energy bill?
The people who build off grid either build because the grid isn't near by, or they simply like the idea of being off grid, and don't care particularly about the time value of money.
You should have heeded the other poster, and tried some calculations.
I disagree. The GPL is better than "no copyrights," because if there were no copyrights, we couldn't easily get the code for binary programs, but with them, and the GPL, we can.
The only issue here is whether this is too expensive to have 100Gig+ bulk memory... but if that's the case then it would be too expensive for a "Solid State Hard Drive" anyway.
Your conclusion paints a false dichotomy, with the unstated assumption being that any form of SSD technology has to at least be as expensive as RAM. This assumption appears unmerited, from direct observation of buyable SSD's today...
Cable Television in many areas of the US is not government sanctioned as a monopoly, but is one anyway, because the situation is indeed natural. Has to do with the economics of digging and competing. Better not to, just tacitly agree to territories.
Interesting little tangle, isn't it? It would seem to me that this sort of trademark shouldn't be granted in the modern, well-connected world, but if you go back in time... well, interesting little conundrum. (Intentionally) using a term that is genuinely generic in one place, but not in another. Makes you think.
I take it that Budweiser means something generic in a foreign country? I can see that. Would be a fair comment, but you have it backwards. In the country in which the Budweiser product was first hosted, the word was/not/ generic. It may be generic elsewhere; that's a different issue. I.e., if the word "Windows" is generic in both its home language and host country, then I would argue the company has no claim to non-generic elsewhere...
Admittedly, these issues of language and culture crossing trade protections are bitches.
That's just a stupid idea. If the word is generic in its own language, and in the country in which it is hosted, it doesn't suddenly become "not generic" elsewhere. If stupid judges in stupid foreign countries want to bamboozle themselves and their countrymen into giving up their rights... well that's a different matter.
Other posters have rightly asked about pacemakers and the like. Be that as it may, there are an entire swath of folks that could be, theoretically, brought down like a device like this regardless of the fact that someone "intent on illegality" could escape it. Remember, there are lots of crimes of passion and irrationality out there. Most of them, really.
Exactly. It was an easy way out, sacrificing hundreds of thousands of civilians to spare US troops (who had volunteered to be used in such a fashion, unlike the civilians who had no such luxury).
Your sentiment here is that the civilians in Japan bore no responsibility whatsoever for their leadership. I disagree.
Slashdot Mirror
There are probably some real damages, in terms of distress and so forth, with hard-to-quantify monetary value. However, if the scam was deliberate, punitive damages are certainly called for.
C//
Re your "correction": there are plenty of
I see what you're saying about sensitivity to data locality. While there is unfortunately as of yet a solution for this, what's wanting here is "transparent storage migration" (on a block level). Think logical volume that is data access frequency aware and likewise aware of drive characteristics. I.e., just make sure the blocks are where you need them most. Not a hard algorithm to write, just needs to be a market need.
While I am indeed sure that the "wanting" thing above will see solutions, I could see how they are unlikely to see the consumer light of day. I'd be more likely to expecting things like hybrid drives featuring something like fusionio technology internally, for a mega cache of sorts. At the right price point, of course. This leverages all sorts of sensible economies of scale, like being able to plug the drive into your favorite RAID controller.
Be that as it may, consider. Is there not some near feature, not so far off day where software itself would be unlikely to be large enough to fill up a drive like this, for many/most users? I think so. Also, there are many classes of work where the job being worked on does fit reasonably well on the drive, and would benefit. Digital image production, various kinds of (small) streaming IO problems and so on.
Anyway, I agree that the probable use of these cards are for cache. I'm just saying that there are plenty of applications that would benefit, should the price point be low enough, and there would certainly be buyers,
In enterprise space, this is obvious from the bevy of buyers who eat up 15K FC/SCSI/SAS drives just to scrape out every last ounce of IOP. Those drives are
I consumer space, it's also obvious. Just look at the enthusiast market.
There's also the workstation space; people with a capital budget, but are sensitive to labor hours.
As far as whether or not their performance claims are lies or not, I can't say. At the moment they are not available, and I'm not at the point of wanting to test prerelease silicon. What I will say is that if they promise "sustained read rates of 800mb/s" and don't deliver in good faith approximately that, I can arrange some very bad mojo for them.
I do feel sure of is that it's possible to do what they say. They really just need to do simple striping of the nonvolatile RAM devices on that card. The way they are doing it has its limits (its scale-limited), but it's also economical. While they could make a 300mB/s SAS form factor, all buyers would then have to buy the RAID card in addition to their drives. Instead, they've married the "RAID" and "drive" circuitry on board. Interesting choice. Says a lot about their intended market.
I think that market is this: enthusiast, and add ons to enterprise storage products (EMC, NetApp, etc). They might seem some workstation buyers also.
C//
And if I had a choice between say, one of the new 64GB SATA flash drives for $50 USD, or a 500GB 7200rpm Seagate SATA HDD for free, I'd go for the Seagate.
There are many buyers who are not like you. The issue is that to many buyers, both $50 and $0 are "free". There is a price point threshold below which the cost is a non expense. This is particularly true in enterprise purchasing situations, where the processing of the paperwork to merely buy and item is hundreds of dollars. While that only assesses the impact of one whole purchase, these and other factors are very real. They are also real in the consumer space, where many classes of buyers routinely behave exactly the same way. I would happily pay $50 for a device that is 8X faster than a hard drive for read and 1000X faster for random access. I already have a 500G hard drive, and it's not full...
I did not imply that you said RAID-0 would solve everything. Rather I am saying that it can be quite difficult to nurse high read/write rates out of RAID structures that don't lose large amounts of data. RAID-5 and RAID-6 won't match the read/write rate of one of fusionio's cards in any configuration from any enterprise seller of storage.
You're right, by the way. I am accepting their read/write rates at face value. This shouldn't trouble you, its a discussion about a theoretical, not-on-the-market product. Let's just suppose.
It might surprise you to learn that we are considering using their product in a clustered storage solution at their listed price. A couple grand per card does not seem out of line to me, as the backing device for a persistent cache. If the product performs as advertised, of course.
I poked around looking for RAID-2 information. Both striping and some sort of error correcting code are mentioned, but otherwise this solution is not discussed much. Tell me more. What are the read/write performance figures for a group of 10-14 drives in this configuration? And what happens if you lose a whole single drive?
C//
14 15K FC won't give you this level of performance, unless you go into RAID-0. I'll discount RAID-0, because it's almost never used in real deployments.
I feel certain that this class of device will appear, and quite soon, in enterprise storage solutions where it will be used as a persistent backing store (cache) in the very RAID arrays that you are talking about. This isn't just guesswork; my position in industry is such that enterprise storage vendors do backflips in order to show me their developing products and roadmaps.
You're losing site of the forest for the trees if you think that the ratio of the costs is the deciding factor for purchase. It's absolutely not. Think it through. If a 64GB device like that were $50, would it bother you to pay for one even if you get (one) 500GB drive for free? The ratio is infinite, but I doubt you'd be bothered, and feel sure that you'd think long and hard about forking over the $50.
You'd think longer and harder if you knew just how hard it is to get reliable, sustained, real-world read rates of 800 MB/s...
C//
http://www.fusionio.com/
Not on the market yet.
Doesn't have the ram, but then given its performance figures you shouldn't care (and if you do, let's not forget you're asking it to do what your OS already does). Same goes for your write-back: at 600MB/s, why?
They're targeting 30$/gb.
C//
That depends, doesn't it?
For example, this device here http://www.fusionio.com/ if affordable enough, would beat any RAID array you're ever likely to put in a workstation. In fact, it's hard to beat a device like this with ANY commercial RAID array. NetApp 3070's can't pull those numbers, and when they manage to get close, you're talking 14 15K FC drives.
Other "features" of your RAID array might be questionable as well. Like the power. And the noise.
One final factor. At SOME price point the fact that there is a ratio difference in price points becomes irrelevant. If you could get a 64GB fusionio card for 20 bucks, you wouldn't really care that a 64GB hard drive is 20 cents. The 100:1 difference in price is nothing, because both prices are nothing. At that scale and price point.
"Consumer price insensitivity" is the term. At some price, the price doesn't matter.
That day WILL arrive.
C//
I'd be interested to hear of people who are running Oracle into things like Lustre. It should work, and work well, but I suspect it will take a brave man to sign it off into production.
The problem with Lustre is that if you configure in the fashion that one usually uses Lustre, there is a small non-zero chance that something can happen to your cluster such that you can realize an availability problem that may impact the entire tier. A future version that does the distributed stripe thing will address that.
(Note that there are still good uses cases for Lustre regardless of the possible tier availability issue, and for the most part the problem can be addressed independently by having each Lustre node be itself HA; note however that in all standard HA patterns, total loss of the storage chassis still spells doom to recent non mirrored data... BTW, if you think storage chassis loss isn't necessary to account in your availability model, then you're golden... depends on what you're doing).
If you're interested in clustered storage ala Acopia, but need it on a SAN, talk to the people at XIV Storage. They're doing clustered storage for block-level data, and are able to easily keep ~40gbit/sec fed. Course, this is a startup. Magna caveat emptor.
C//
All good thoughts. Agree and all that. This being slashdot, instead of talking at length about agreement we'll have to up the ante and disagree about something (hah). Let's talk about that "you still need to mirror the SSD until your insane" comment. I'm finding in-box redundancy to be less and less useful. Check out http://www.isilon.com./ These guys, and emerging vendors like them, are beginning to move away from in-array replication of data to cross-array/cross-box replication. Packets are striped across the filers, instead of inside of the array. The beauty here, is that you can actually lose the filer chassis entirely, and you don't lose any data. You can replace when convenient, and get closer to the idea of "ignoring the chance of RAID-5/RAID-6 simultaneous disk failure". It's still not entirely there, mind, but since RAID rebuilds involve any disk in all of your racks, you have no need to do anything, the system is essentially self-healing.
Soon, Lustre file system will have such a feature. I can't wait.
Another vendor, mostly still in stealth mode, is here: http://www.xivstorage.com/
I'm sure you will find more and more turnkey cluster storage vendors hitting the market soon, precisely because clustered storage solves bandwidth, capacity, and manageability problems so well. All sorts of serviceability problems are cleanly solved by breaking away from classic RAID and moving to the right kind of cluster architecture.... problem is, until recently it's been a science project. Turnkey cluster appliances are right around the corner (or here, if you can live within Isilon's NAS-only constraints).
C//
It's not all one way or the other. At some price point, there are consumers who will buy the fast thing, just to have SOME of the fast thing. One is not required to have only one thing inside one's computer. For example, I could make productive use of a 32 or 64GB flash system if it were fast enough and cheap enough RIGHT NOW, regardless of the fact that I have a 500GB SATA drive in my system in order to store movies and images.
There are also valuable business applications for the same technology. If 64GB flash were "relatively affordable" and noticeably faster and more effective than a good RAID array, then these flash drives could be an important component of an enterprise storage system. Think "cache times ten, plus non volatile."
I'd love to build a Lustre Cluster where each of the nodes had one of these as a persistent backing store for cache: http://www.fusionio.com./ As an added plus, their stated market price of $30/GB is ALMOST worth it.
C//
Now you're merely lying.
An investment of equal size has a return of it own. That return COMPOUNDS. A mutual fund will average about 11.5% annually, COMPOUNDED, when you pick a simple S&P 500 index.
You have NOT performed the calculations. You've only performed the half of them. You calculated, on the back of a napkin, so to speak, how much money you would "save." You did not calculate, however, how much money you would fail to gain by not investing that money in good investments elsewhere. Not to mention: with the "good investments elsewhere," your principal is still in your hands at the end. With your current strategy, the money is not.
A quick shot in excel shows me this. $35,000 at 11.5% interest is worth $100K after 10 years. Did your 10 year calculation show a similar return on energy bill?
I KNOW THAT IT DID NOT.
C//
You're not one of them, and you haven't performed any calculations.
It's easy. But no.
PERFORM THE GODDAMN CALCULATIONS.
C//
The people who build off grid either build because the grid isn't near by, or they simply like the idea of being off grid, and don't care particularly about the time value of money.
You should have heeded the other poster, and tried some calculations.
Truly.
C//
I disagree. The GPL is better than "no copyrights," because if there were no copyrights, we couldn't easily get the code for binary programs, but with them, and the GPL, we can.
C//
Oh. Rereading, I see I misunderstood. But according to Slashdot, I'm 'Insightful'.
*chortle*
C//
The only issue here is whether this is too expensive to have 100Gig+ bulk memory... but if that's the case then it would be too expensive for a "Solid State Hard Drive" anyway.
Your conclusion paints a false dichotomy, with the unstated assumption being that any form of SSD technology has to at least be as expensive as RAM. This assumption appears unmerited, from direct observation of buyable SSD's today...
C//
Cable Television in many areas of the US is not government sanctioned as a monopoly, but is one anyway, because the situation is indeed natural. Has to do with the economics of digging and competing. Better not to, just tacitly agree to territories.
C//
Interesting little tangle, isn't it? It would seem to me that this sort of trademark shouldn't be granted in the modern, well-connected world, but if you go back in time... well, interesting little conundrum. (Intentionally) using a term that is genuinely generic in one place, but not in another. Makes you think.
What you are missing is a bevy of people calling themselves scientists who are instead trapped deeply in their own mythology.
C//
I take it that Budweiser means something generic in a foreign country? I can see that. Would be a fair comment, but you have it backwards. In the country in which the Budweiser product was first hosted, the word was /not/ generic. It may be generic elsewhere; that's a different issue. I.e., if the word "Windows" is generic in both its home language and host country, then I would argue the company has no claim to non-generic elsewhere...
Admittedly, these issues of language and culture crossing trade protections are bitches.
C//
That's just a stupid idea. If the word is generic in its own language, and in the country in which it is hosted, it doesn't suddenly become "not generic" elsewhere. If stupid judges in stupid foreign countries want to bamboozle themselves and their countrymen into giving up their rights... well that's a different matter.
C//
The product you are looking for is called the "ram san." Texas Memory Systems makes it.
C//
Have to get your tailhook into the right position and elevation. Check. :)
C//
Other posters have rightly asked about pacemakers and the like. Be that as it may, there are an entire swath of folks that could be, theoretically, brought down like a device like this regardless of the fact that someone "intent on illegality" could escape it. Remember, there are lots of crimes of passion and irrationality out there. Most of them, really.
C//
Exactly. It was an easy way out, sacrificing hundreds of thousands of civilians to spare US troops (who had volunteered to be used in such a fashion, unlike the civilians who had no such luxury).
Your sentiment here is that the civilians in Japan bore no responsibility whatsoever for their leadership. I disagree.
C//