I hope it eventually becomes "fast enough to settle on one standard". It's getting to the point where if the software support was there we could just settle on a standard like USB3.1 gen 2 (the 10 Gbit one) for internal disks and external peripherals with some kind of PCIe slotted flash solution for people who wanted stupid fast speed that only shows on benchmarks.
Maybe by the next major revision they will figure out how to come up with a way to unify interface standards. The bus speed increases are making it increasingly about who has the best connector and longest usable cable lengths, not who is 1-2 GBits/sec faster. What's crazy is that a SSD on USB3.1 is probably capable of the kind of disk I/O only the biggest 8 gig fiber channel SANs with spinning disk could deliver just five years ago.
I have a 10 year old son and as much as I hate alarmism, I do find the allure of technology kind of scary.
We give our son "screen time" (PC, XBox or iPad) but we usually limit it to an hour per day. But if given the ability, he would play much more than that. It's like a compulsion. And it's often a struggle when his hour is up to get him to quit.
When we go places, I see lots of younger kids absolutely glued to a screen (iPad, iPhone usually). The touchscreen devices seem to have some kind of extra allure, which I associate with the fact that they have a tactile component different than a game controller or keyboard/mouse.
The Romans had little use for unifying the refugees of their occupied lands. They killed all that resisted and the rest were enslaved. Those that remained assimilted Roman culture because the opposite was death. This basic structure is the same in every place where force was successful.
The West's military misadventures have failed because Western militaries have become preoccupied with defeating armies and weapon systems. They are no longer focused on defeating a nation. Military force can only be successful to the extent that it is willing to defeat a people and break not only their ability to fight but to their willingness to fight.
This is not done through "nation building", nurturing or any other touchy-feely behavior. It is done by killing people who resist and destroying their places of living. Every act of resistence should be dealt with death and destruction until everyone willing to fight is dead and everyone else won't fight.
You approach a village and you take fire from it? You level the village and kill everyone who resists. You keep doing this and you will not have any resistence. People will learn that resistence is futile, that resistence means death.
There is nothing nice about this. It is vicious and it is brutal. Which is why we should not engage in military actions unless we are willing to pursue it. Because the opposite is viciousness and brutality for both sides without any resolution.
It worked in Gaul, Carthage, post-WW I Russia, the Native Americans, Tiananmen Square, Nazi Germany (against the Nazis), Eastern Georgia under German Sherman, Imperial Japan...
...instead of enabling or encouraging farmers to become water speculators?
If the inputs are priced more accurately than the outputs should reflect these costs. If almonds take a lot of water to grow, then almonds should be more expensive to reflect the higher water prices.
Allowing farmers to sell unused water seems like an invitation for speculators to buy farms not for the purpose of farming but to just speculate in water, or worse, figure ways to manipulate both commodity markets and water supplies.
A better solution might be encouraging water CREATION through incentives for water recycling or desalination through renewable energy.
Also, despite the whining, it is a fine OS. It's only real issue is the start screen is inefficient to us. Not impossible, not insurmountable, just inefficient. You can use a system with it just fine. What's more, it is a real easy problem to fix. Buy Start 8, or get Classic Shell for free and you're done, a classic start menu that works nice.
When I first saw this topic my gut reaction was "Those bastards!" and then I remembered I've been running Win 8 for the last year on a Surface Pro with Classic Shell without any major annoyances other than some of the built-in stuff that Classic Shell doesn't change.
And then I started thinking, where would MS be if they hadn't fucked with the UI on Win8? That seems to be their biggest problem, not the OS changes itself.
Well, I think most hazardous substances that exist in nature usually exist as some kind of stable mineral combination, not as a pure element in concentration.
Mercury is found as cinnabar in nature, not as the liquid refined metal. The problem with the pure metal is that once in nature it combines with other shit or concentrates in the food chain, which is why its kind of a bad idea.
"Ridiculous money" is sort of a relative. I'd argue that Tag customers pay ridiculous money relative to their income, since most Tag customers are merely sort of affluent, not "rich".
Ultimately I think one solution would be to require the manufacturer to take back and arrange for recycling or proper disposal.
Which would tack on how much to the cost of CFLs? And involve how much carbon emission in the pickup and movement of used CFLs?
The better solution is to not push ZOMG! GREEN! products which have a bunch of hazardous waste in them.
I have no idea how many CFLs end up in landfills, but I do sometimes wonder if the environmental cost of that much mercury in the environment is worse than the CO2 emissions prevented by using them.
I asked my last Uber driver (who was driving her BMW 328i) how she liked it, the money, etc. Her answer was "I drove for a week and quit my two other jobs. It's all I do now."
Now, if you did some kind of spreadsheet and factored in maintenance, correct insurance, etc, it might end up being "not so good of a deal" but that's really impossible to say unless you're the driver.
Calling it "exploitation" is hyperbolic in the same way that a Marxist calls anyone working for a capitalist "exploited". OK, within a specific analytic framework and with a specific set of value judgements made maybe it is, but at the same time you can find a lot of other people who one group calls "exploited" who say "What? I'm totally satisfied with this arrangement."
I think even in some jurisdictions there is a temptation to collect data on citation sources and create allocation formulas for state grants to LEOs based on citation data.
"Everybody" wins -- the system seems to be free of conflicts of interest, the legislature gets to keep the LEOs happy. Except citizens, I guess.
You use the Stingray to build evidence, you then construct a case so that what you find out via Stingray can be presented as having been discovered legally.
That way defense counsel has nothing to challenge and the secret/illegal intelligence gathering stays safely hidden.
The appropriate literary reference isn't some John Grisham novel, it's Franz Kafka.
BTW I have not had a speeding ticket in over a decade and I do not park in handicapped spots but the use of law enforcement for generating revenue is a terrible trend and needs to be stopped.
The way to do this is to remove the reward incentive from fines.
If the fines go to the general fund for whatever the highest political entity is (eg, if Ferguson, MO fines somebody, the money goes to the state's general fund) then you remove any financial incentive for the local police to fine anyone, because the money leaves their jurisdiction.
Almost all of the "policing for profit" problems seem to exist because the cops, or someone closely supervising the cops, gets the money. This creates an incentive for them to do more of the same because they get paid.
If the money gets distributed into a larger pool, they have no incentive to collect fines outside of whatever basic incentives the police have to pursue safety.
I'm sure there will be the usual complaints about why the police or local government should collect money for labor-intensive enforcement efforts or why the police won't bother to enforce some laws.
Empires have grown and fallen by doing exactly, morally, what the nazis did. But never mechanically what they did. We've destroyed peoples, but not with the efficiency and clarity of the Holocaust.
I think some of the Roman conquests might actually have matched the Nazis. Carthage was largely exterminated completely. Julius Ceasar's Gallic campaign certainly encountered some smaller opponents who resisted and the end result was anyone not dead was sold into slavery.
Parts of the 30 Years' war were pretty intensive, too, although it's a debate as to how much of this primary and how much of this was a secondary due to famine and disruption.
Past empires may have lacked the kind of industrial metaphor of the Final Solution, but that's maybe because past empires lacked some of the concepts about industrial processes and organization.
a BS proposition, since you're streaming from the internet to one device, then locally restreaming to the AppleTV -- way to clog your network.
I need to get some monitoring going and test various airplay options, because I have a suspicion that where it can, the device may pass off the stream URL to the AppleTV so it streams natively. Obviously, in some cases it can't and this is where I think the difference between airplay and airplay mirroring might come in.
Overall, I agree with the notion that it's kind of bogus solution and about the only time I use it with any intensity is browing YouTube on an iPad but playing videos on the TV. The browsing is simply better on the iPad.
But, on the other hand, even if it is a kludge, it's a pretty great kludge because it works. I have my ATV on a wired gigabit connection and the access point iDevices connect to is 2.4/5Ghz N, so network performance hasn't been an issue. I could see where it could be an issue if it was all wifi.
I can only think "run apps" but you can mostly do that with mirroring which eliminates a whole lot of ugly compatibility issues (for Apple, and app developers) even if the experience is lacking in some ways. Even with apps, the whole interactive controller issue gets kind of weird if a phone, tablet or the remote doesn't make sense.
By and large, Apple TV does what its supposed to do -- play media on your TV from your phone, tablet, PC or streaming.
The lack of an Amazon Instant app is annoying, but Netflix is there. The rest of the content is kind of ho-hum, but then again, mirroring and airplay solves some of those issues, like Amazon Instant or many other media playback apps.
I think a lot of people would like the Apple TV to be "more" of something, but so many of those things are a weird fit or hinge on other lateral expansions like more input devices or something else or end up overlapping with what it does now but without a significant expansion of functionality.
Not sure why you're modded down (well, OK, this is Slashdot so I have some idea..) but I think everything you've said is pretty reasonable.
I think sometimes it's not *what* Apple does but *how* they do it. My experience with Lightning is that it's been a great connector, far better than micro-USB, with it's mandatory 3x orientation changes to get something plugged in.
What was obnoxious was how it was introduced -- with the iPhone 5, with few adapters or any third-party cables available at retail. This introduced a ton of immediate compatibility problems on a device that really should have immediate accessory availability. They could have introduced it with iPad3 prior to iPhone 5 and given some time for more cables, more adapters, etc. to be available for the phone release with the new connector. iPad connectivity is far less of an issue, IMHO, than iPhone and would have been less disruptive.
The other thing that seemed annoying was the level of control being exercised over cables, which greatly slowed the creation of third party cables/connectors/devices. I don't know what the long-term impact of this was, really, but short term it was obnoxious and limited accessory availability. I think they're also a little control-freakish when it comes to what you can connect to it -- I think a slightly more open stance would be valuable.
Lightning was a win as a connector, but the process was stupid and user hostile. Much of it doesn't seem to matter in the long run, and maybe some of this process provides hard-to-see value to end users in terms of quality or compatibility, but I can see where this could be done a lot smoother.
Yeah, but what percentage of people never go outdoors ever? And how healthy and mentally well-balanced are those people?
There's a lot of people in cold northern climates that almost never go outside in the winter months. -20F with a 20 MPH wind may not be as completely hostile as Mars, but you pretty much have to have all your exposed skin covered and wear heavy insulated clothing so you might as well be wearing a spacesuit.
Sure, there's a small percentage of people who are athletic and go outside in that kind of cold anyway, but they're a very small minority.
I live in Minnesota and I know people who have heated, attached garages, park their cars at work in heated underground parking and Monday through Friday never go outside. They often only go outside when they absolutely have to, and even then its a dash from car to building and vice-versa.
They often seem a lot happier than people who ride the bus and have to stand outside for 20 minutes or longer to catch a bus and then walk 4-5 blocks in the cold, especially when it gets below zero. That kind of thing is a total grind and very wearying.
I think they should re-run the test again, this time with 5 disks from the "best" of the disks from the torture test to see how much variability there is within a single drive model. It wouldn't be much for real statistics, but it should give you an idea how much variation you might expect from a single model.
Then I'd like to do it all over again and put the disks into an array and see how this affects disk lifetime. Do disks fail sooner than expected when further abused with array overhead? Perhaps last longer as the array will distribute writes across all disks, allowing basically the array to last for N-1 * disks lifetime?
The latter part really interests me considering what write-intensive SSDs cost in enterprise SANs. It might be that perhaps with the workload spread across enough commodity SSDs their durability is less of an issue (provided hotspares, good load distribution, double or triple parity, etc).
It may actually prove economically viable to build arrays with cheap SSDs if over the lifetime of the array disk replacement costs don't exceed the upfront cost of write intensive SSDs and the reliability of the whole array is good enough that failures don't take out the array or require excessive maintenance.
I still think there's so much performance advantage to be gained from the OS and apps on the SSD that the only real purpose of spinning rust is capacity and whatever reliability it provides over SSDs. The torture test seems to indicate that the reliability factor isn't that much to worry about.
I think what you'd really want is something where the SSD takes all writes, mirrors them to HDDs and caches all reads to SSD, but can read AND write to the HDDs if there is a loss of SSDs.
Bonus points for actual SAN-like behavior, where the total system capacity is actually measured by the HDD capacity and the system is capable of sane behavior, like redirecting writes to HDD if the SSD write cache overflows and preserving some portion of high-count read cache blocks so that unusually large reads don't destroy the read cache.
Those Seagate hybrid drives came kind of close, but never had enough flash to really accomplish this. I think there are SATA-capable array cards capable of building hybrid arrays read-biased to SSDs. There are probably filesystem methods for doing this, too.
I've owned Tivos since 2002 and I've only had one blow a drive, a series 3 I bought from WeakKnees with an upgraded disk in it. The drive didn't fail spectacularly or even completely, we just had a ton of playback problems and recordings that grew increasingly unreliable. That Tivo was bought in 2007 and the drive was replaced last fall.
The original Tivo I bought in 2002 finally got tossed without a drive failure when Comcast gave up on analog SD channels a couple of years ago. I think this was after broadcast went digital, making it useless. I had ideas of adding one of those boxes for people with analog-only TVs, but I wasn't sure if the IR emitter would work with the digital TV converter and it was kind of silly to jump through those hoops to record HD as SD only to watch it on an HD TV.
The other two Series 3s have original drives and work fine.
Assuming that their customers do not encrypt their data before they place it on DropBox (which would make sense because DropBox customers are rarely CIOs themselves), then DropBox is most likely hashing the content and only storing a single copy of a file even if there are thousand virtual instances of that same file throughout their system.
Wouldn't it make more sense for them to dedupe on some kind of variable block size level than the file level? If someone uploads v1 of some 25 meg powerpoint file and then version 2 with just a single page changed, you can't dudupe anything. If you did it at the block level you'd be able to dedupe much more.
And I would wager that kind of syndrome is common, with someone who works on project X and has tons of files with identical content embedded in them. Plus it would work with encrypted data as well -- you don't care what the data is, just that some chunks of it happened to be identical.
Slashdot Mirror
I hope it eventually becomes "fast enough to settle on one standard". It's getting to the point where if the software support was there we could just settle on a standard like USB3.1 gen 2 (the 10 Gbit one) for internal disks and external peripherals with some kind of PCIe slotted flash solution for people who wanted stupid fast speed that only shows on benchmarks.
Maybe by the next major revision they will figure out how to come up with a way to unify interface standards. The bus speed increases are making it increasingly about who has the best connector and longest usable cable lengths, not who is 1-2 GBits/sec faster. What's crazy is that a SSD on USB3.1 is probably capable of the kind of disk I/O only the biggest 8 gig fiber channel SANs with spinning disk could deliver just five years ago.
That's the problem, though, the allure overpowers any sense of self-control.
I have a 10 year old son and as much as I hate alarmism, I do find the allure of technology kind of scary.
We give our son "screen time" (PC, XBox or iPad) but we usually limit it to an hour per day. But if given the ability, he would play much more than that. It's like a compulsion. And it's often a struggle when his hour is up to get him to quit.
When we go places, I see lots of younger kids absolutely glued to a screen (iPad, iPhone usually). The touchscreen devices seem to have some kind of extra allure, which I associate with the fact that they have a tactile component different than a game controller or keyboard/mouse.
The Romans had little use for unifying the refugees of their occupied lands. They killed all that resisted and the rest were enslaved. Those that remained assimilted Roman culture because the opposite was death. This basic structure is the same in every place where force was successful.
The West's military misadventures have failed because Western militaries have become preoccupied with defeating armies and weapon systems. They are no longer focused on defeating a nation. Military force can only be successful to the extent that it is willing to defeat a people and break not only their ability to fight but to their willingness to fight.
This is not done through "nation building", nurturing or any other touchy-feely behavior. It is done by killing people who resist and destroying their places of living. Every act of resistence should be dealt with death and destruction until everyone willing to fight is dead and everyone else won't fight.
You approach a village and you take fire from it? You level the village and kill everyone who resists. You keep doing this and you will not have any resistence. People will learn that resistence is futile, that resistence means death.
There is nothing nice about this. It is vicious and it is brutal. Which is why we should not engage in military actions unless we are willing to pursue it. Because the opposite is viciousness and brutality for both sides without any resolution.
It worked in Gaul, Carthage, post-WW I Russia, the Native Americans, Tiananmen Square, Nazi Germany (against the Nazis), Eastern Georgia under German Sherman, Imperial Japan...
...instead of enabling or encouraging farmers to become water speculators?
If the inputs are priced more accurately than the outputs should reflect these costs. If almonds take a lot of water to grow, then almonds should be more expensive to reflect the higher water prices.
Allowing farmers to sell unused water seems like an invitation for speculators to buy farms not for the purpose of farming but to just speculate in water, or worse, figure ways to manipulate both commodity markets and water supplies.
A better solution might be encouraging water CREATION through incentives for water recycling or desalination through renewable energy.
Also, despite the whining, it is a fine OS. It's only real issue is the start screen is inefficient to us. Not impossible, not insurmountable, just inefficient. You can use a system with it just fine. What's more, it is a real easy problem to fix. Buy Start 8, or get Classic Shell for free and you're done, a classic start menu that works nice.
When I first saw this topic my gut reaction was "Those bastards!" and then I remembered I've been running Win 8 for the last year on a Surface Pro with Classic Shell without any major annoyances other than some of the built-in stuff that Classic Shell doesn't change.
And then I started thinking, where would MS be if they hadn't fucked with the UI on Win8? That seems to be their biggest problem, not the OS changes itself.
Well, I think most hazardous substances that exist in nature usually exist as some kind of stable mineral combination, not as a pure element in concentration.
Mercury is found as cinnabar in nature, not as the liquid refined metal. The problem with the pure metal is that once in nature it combines with other shit or concentrates in the food chain, which is why its kind of a bad idea.
"Ridiculous money" is sort of a relative. I'd argue that Tag customers pay ridiculous money relative to their income, since most Tag customers are merely sort of affluent, not "rich".
Ultimately I think one solution would be to require the manufacturer to take back and arrange for recycling or proper disposal.
Which would tack on how much to the cost of CFLs? And involve how much carbon emission in the pickup and movement of used CFLs?
The better solution is to not push ZOMG! GREEN! products which have a bunch of hazardous waste in them.
I have no idea how many CFLs end up in landfills, but I do sometimes wonder if the environmental cost of that much mercury in the environment is worse than the CO2 emissions prevented by using them.
I asked my last Uber driver (who was driving her BMW 328i) how she liked it, the money, etc. Her answer was "I drove for a week and quit my two other jobs. It's all I do now."
Now, if you did some kind of spreadsheet and factored in maintenance, correct insurance, etc, it might end up being "not so good of a deal" but that's really impossible to say unless you're the driver.
Calling it "exploitation" is hyperbolic in the same way that a Marxist calls anyone working for a capitalist "exploited". OK, within a specific analytic framework and with a specific set of value judgements made maybe it is, but at the same time you can find a lot of other people who one group calls "exploited" who say "What? I'm totally satisfied with this arrangement."
I think even in some jurisdictions there is a temptation to collect data on citation sources and create allocation formulas for state grants to LEOs based on citation data.
"Everybody" wins -- the system seems to be free of conflicts of interest, the legislature gets to keep the LEOs happy. Except citizens, I guess.
You could use your own CA and generate self signed certificates.
You use the Stingray to build evidence, you then construct a case so that what you find out via Stingray can be presented as having been discovered legally.
That way defense counsel has nothing to challenge and the secret/illegal intelligence gathering stays safely hidden.
The appropriate literary reference isn't some John Grisham novel, it's Franz Kafka.
BTW I have not had a speeding ticket in over a decade and I do not park in handicapped spots but the use of law enforcement for generating revenue is a terrible trend and needs to be stopped.
The way to do this is to remove the reward incentive from fines.
If the fines go to the general fund for whatever the highest political entity is (eg, if Ferguson, MO fines somebody, the money goes to the state's general fund) then you remove any financial incentive for the local police to fine anyone, because the money leaves their jurisdiction.
Almost all of the "policing for profit" problems seem to exist because the cops, or someone closely supervising the cops, gets the money. This creates an incentive for them to do more of the same because they get paid.
If the money gets distributed into a larger pool, they have no incentive to collect fines outside of whatever basic incentives the police have to pursue safety.
I'm sure there will be the usual complaints about why the police or local government should collect money for labor-intensive enforcement efforts or why the police won't bother to enforce some laws.
Empires have grown and fallen by doing exactly, morally, what the nazis did. But never mechanically what they did. We've destroyed peoples, but not with the efficiency and clarity of the Holocaust.
I think some of the Roman conquests might actually have matched the Nazis. Carthage was largely exterminated completely. Julius Ceasar's Gallic campaign certainly encountered some smaller opponents who resisted and the end result was anyone not dead was sold into slavery.
Parts of the 30 Years' war were pretty intensive, too, although it's a debate as to how much of this primary and how much of this was a secondary due to famine and disruption.
Past empires may have lacked the kind of industrial metaphor of the Final Solution, but that's maybe because past empires lacked some of the concepts about industrial processes and organization.
a BS proposition, since you're streaming from the internet to one device, then locally restreaming to the AppleTV -- way to clog your network.
I need to get some monitoring going and test various airplay options, because I have a suspicion that where it can, the device may pass off the stream URL to the AppleTV so it streams natively. Obviously, in some cases it can't and this is where I think the difference between airplay and airplay mirroring might come in.
Overall, I agree with the notion that it's kind of bogus solution and about the only time I use it with any intensity is browing YouTube on an iPad but playing videos on the TV. The browsing is simply better on the iPad.
But, on the other hand, even if it is a kludge, it's a pretty great kludge because it works. I have my ATV on a wired gigabit connection and the access point iDevices connect to is 2.4/5Ghz N, so network performance hasn't been an issue. I could see where it could be an issue if it was all wifi.
I'm curious what more it's supposed to do.
I can only think "run apps" but you can mostly do that with mirroring which eliminates a whole lot of ugly compatibility issues (for Apple, and app developers) even if the experience is lacking in some ways. Even with apps, the whole interactive controller issue gets kind of weird if a phone, tablet or the remote doesn't make sense.
By and large, Apple TV does what its supposed to do -- play media on your TV from your phone, tablet, PC or streaming.
The lack of an Amazon Instant app is annoying, but Netflix is there. The rest of the content is kind of ho-hum, but then again, mirroring and airplay solves some of those issues, like Amazon Instant or many other media playback apps.
I think a lot of people would like the Apple TV to be "more" of something, but so many of those things are a weird fit or hinge on other lateral expansions like more input devices or something else or end up overlapping with what it does now but without a significant expansion of functionality.
Not sure why you're modded down (well, OK, this is Slashdot so I have some idea..) but I think everything you've said is pretty reasonable.
I think sometimes it's not *what* Apple does but *how* they do it. My experience with Lightning is that it's been a great connector, far better than micro-USB, with it's mandatory 3x orientation changes to get something plugged in.
What was obnoxious was how it was introduced -- with the iPhone 5, with few adapters or any third-party cables available at retail. This introduced a ton of immediate compatibility problems on a device that really should have immediate accessory availability. They could have introduced it with iPad3 prior to iPhone 5 and given some time for more cables, more adapters, etc. to be available for the phone release with the new connector. iPad connectivity is far less of an issue, IMHO, than iPhone and would have been less disruptive.
The other thing that seemed annoying was the level of control being exercised over cables, which greatly slowed the creation of third party cables/connectors/devices. I don't know what the long-term impact of this was, really, but short term it was obnoxious and limited accessory availability. I think they're also a little control-freakish when it comes to what you can connect to it -- I think a slightly more open stance would be valuable.
Lightning was a win as a connector, but the process was stupid and user hostile. Much of it doesn't seem to matter in the long run, and maybe some of this process provides hard-to-see value to end users in terms of quality or compatibility, but I can see where this could be done a lot smoother.
Yeah, but what percentage of people never go outdoors ever? And how healthy and mentally well-balanced are those people?
There's a lot of people in cold northern climates that almost never go outside in the winter months. -20F with a 20 MPH wind may not be as completely hostile as Mars, but you pretty much have to have all your exposed skin covered and wear heavy insulated clothing so you might as well be wearing a spacesuit.
Sure, there's a small percentage of people who are athletic and go outside in that kind of cold anyway, but they're a very small minority.
I live in Minnesota and I know people who have heated, attached garages, park their cars at work in heated underground parking and Monday through Friday never go outside. They often only go outside when they absolutely have to, and even then its a dash from car to building and vice-versa.
They often seem a lot happier than people who ride the bus and have to stand outside for 20 minutes or longer to catch a bus and then walk 4-5 blocks in the cold, especially when it gets below zero. That kind of thing is a total grind and very wearying.
I think they should re-run the test again, this time with 5 disks from the "best" of the disks from the torture test to see how much variability there is within a single drive model. It wouldn't be much for real statistics, but it should give you an idea how much variation you might expect from a single model.
Then I'd like to do it all over again and put the disks into an array and see how this affects disk lifetime. Do disks fail sooner than expected when further abused with array overhead? Perhaps last longer as the array will distribute writes across all disks, allowing basically the array to last for N-1 * disks lifetime?
The latter part really interests me considering what write-intensive SSDs cost in enterprise SANs. It might be that perhaps with the workload spread across enough commodity SSDs their durability is less of an issue (provided hotspares, good load distribution, double or triple parity, etc).
It may actually prove economically viable to build arrays with cheap SSDs if over the lifetime of the array disk replacement costs don't exceed the upfront cost of write intensive SSDs and the reliability of the whole array is good enough that failures don't take out the array or require excessive maintenance.
I still think there's so much performance advantage to be gained from the OS and apps on the SSD that the only real purpose of spinning rust is capacity and whatever reliability it provides over SSDs. The torture test seems to indicate that the reliability factor isn't that much to worry about.
I think what you'd really want is something where the SSD takes all writes, mirrors them to HDDs and caches all reads to SSD, but can read AND write to the HDDs if there is a loss of SSDs.
Bonus points for actual SAN-like behavior, where the total system capacity is actually measured by the HDD capacity and the system is capable of sane behavior, like redirecting writes to HDD if the SSD write cache overflows and preserving some portion of high-count read cache blocks so that unusually large reads don't destroy the read cache.
Those Seagate hybrid drives came kind of close, but never had enough flash to really accomplish this. I think there are SATA-capable array cards capable of building hybrid arrays read-biased to SSDs. There are probably filesystem methods for doing this, too.
I've owned Tivos since 2002 and I've only had one blow a drive, a series 3 I bought from WeakKnees with an upgraded disk in it. The drive didn't fail spectacularly or even completely, we just had a ton of playback problems and recordings that grew increasingly unreliable. That Tivo was bought in 2007 and the drive was replaced last fall.
The original Tivo I bought in 2002 finally got tossed without a drive failure when Comcast gave up on analog SD channels a couple of years ago. I think this was after broadcast went digital, making it useless. I had ideas of adding one of those boxes for people with analog-only TVs, but I wasn't sure if the IR emitter would work with the digital TV converter and it was kind of silly to jump through those hoops to record HD as SD only to watch it on an HD TV.
The other two Series 3s have original drives and work fine.
Assuming that their customers do not encrypt their data before they place it on DropBox (which would make sense because DropBox customers are rarely CIOs themselves), then DropBox is most likely hashing the content and only storing a single copy of a file even if there are thousand virtual instances of that same file throughout their system.
Wouldn't it make more sense for them to dedupe on some kind of variable block size level than the file level? If someone uploads v1 of some 25 meg powerpoint file and then version 2 with just a single page changed, you can't dudupe anything. If you did it at the block level you'd be able to dedupe much more.
And I would wager that kind of syndrome is common, with someone who works on project X and has tons of files with identical content embedded in them. Plus it would work with encrypted data as well -- you don't care what the data is, just that some chunks of it happened to be identical.