Could they use the heated containment water to drive a Stirling engine which would pump more water?
I'm not a nuclear scientist (let alone a rocket scientist), but somehow it seems like there's a way to (relatively simply) use the heat from containment cooling to pump water.
And is there a reason the containment water couldn't be a loop with a cooling stage so it could be self-replenishing? It seems to make more sense if you consider the idea of the containment heat being used to drive pumps which circulated the water.
I'm sure there are good reasons why this is a Rube Goldberg perpetual motion kind of an idea, but it seems like a way where as long as it generates heat it could pump its own cooling.
It would seem to be pretty trivial to implement a feature where the phone is aware of the account's data plan details (used, available, etc).
The phone could then have some user configurable warning threshold when the plan's available data goes below that threshold.
Apps and features (like wifi-assist) could then have a setting that allow/prohibits them to use data when the threshold has been exceeded.
If there was some standardized method of obtaining and communicating this info, shared data plans could manage these settings on the carrier account page per device -- individual caps, thresholds relative to individual caps or total shared caps.
This sounds like an application of a de minimis principal -- arguing that the scale of Wikimedia's traffic isn't large enough to give it standing.
I seem to remember learning this phrase in college in a political science class about some guy who sued the government to get some kind of information on defense appropriations, but was rejected because his individual contribution was so small that it essentially didn't matter.
I sometimes wonder if the answer is something we don't want it to be -- that heavy handed policing is repugnant, but the nature of our society is that the poor, urban populations suffer from so much social malaise that without heavy handed policing they will tear themselves apart while inflicting collateral damage on the rest of society through crime and violence?
The causes of malaise are often unjust -- discrimination, lack of opportunity, but also include self-inflicted problems of unplanned/bad parenting, purposeful rejection of positive social choices (dropping out of school, etc) and so on. It's not completely their fault, but it's not completely a question of inescapable victimhood, either.
Heavy-handed policing is likely not the best course to *solve* these problems, but the scale and nature of them is such that the costs and scale of the social welfare solutions which could possibly be more effective are seen as an impossible burden (which itself is a byproduct of economic inequality and ineffective governance).
Heavy-handed policing is thus seen as the most obtainable possible solution -- the least worst possible alternative.
In other words, we may hate heavy handed policing because it also as collateral damage of all kinds, including gross injustices, but without it we may have a kind of chaos that ends up being worse.
All of this just seems to reveal the genius that was expressed in John Browning's.45 auto 1911. Grip size, grip angle, cartridge effectiveness, felt recoil, accuracy -- it's all so much greater than the sum of it parts.
And despite the "advancements" of other rounds and gun designs, I have yet to shoot an automatic that doesn't feel like something got sacrificed for whatever advancement a specific gun brought to the table. All double-stacks seem to compromise ergonomics in some way through larger grips, and most of the "other" cartridges try to compensate for smaller bullet cross section and reduced case capacity with high-pressure loads in an attempt to throw their smaller, lighter projectiles fast enough to get sufficient energy yet rely on highly engineered projectiles to prevent over-penetration.
I have large enough hands and enough body mass that some of these advancements are less of a trade-off for me personally (I have carried a Glock 29 for years), but it's still patently obvious just how wonderful the 1911 is.
The only thing I'm not a huge fan of is carrying a 1911 cocked and locked. I believe the design, but personal hesitancy makes it a non-starter for me. It doesn't help that when I was first introduced to it by a more experienced shooter and questioned the safety that the gun being demonstrated actually managed to drop the hammer without the grip safety (a problem with that specific gun's parts). The gun was unloaded and there were no consequence, but you might equate it to bad experience with a dog when you're young.
If you're not searching for a specific brand name of something, you will often get 87 nearly identical products (often with wildly varying ratings, chock full o'astroturfed reviews) for junk made in China.
Half the time I search on Amazon I feel like I'm just getting an iframe with the results from Alibaba.
Forgetting the politics, economics and global events surrounding those $25 trillion, if we *had* decided to spend $25 trillion (which is nearly 2 years of total US GDP) on fusion would we have even a small scale utility version by now?
I'm assuming that given the size of that amount of money and the actual US GDP, it would take a massive effort just to practically even spend that kind of money, let alone perform the science and engineering it might enable (such as building facilities, etc).
The Samsung stick has only 20% of the speed of DDR3 and enough extra latency that it would be a painful step backward.
Although isn't that basically the system used by smartphones and tablets? Suspend the app and page it to flash when you switch to another context? But they also have much smaller applications and have like 1-2 GM RAM to begin with.
I think where that system might work better overall is on virtualization hosts where active memory is much lower than assigned memory. Those inactive pages available on a fast swap drive would make high density hosts a lot less painful than generic disk swapping.
The FBI had standardized on.38 special and got their asses kicked in the infamous "Miami shootout".
The FBI turned to S&W who worked with Norma to develop the 10mm round which got put in the S&W 10xx series of autoloaders, sharing the same frame as the 45xx series of.45s.
The full-power Norma 10mm was hella stopping power, but it was unpopular with most women and some men due to size and recoil. The original Norma loads were close to.41 Magnum power. I handload both cartridges and believe this is pretty true -- my 610 with maxed-out 10mms feel about the same as my Model 57 with middling loads.
The FBI decided they didn't like the 10mm afterall, but some bright bulb at S&W realized that if you cut down the 10mm case by a few mm you could have a grip size and recoil similar to 9mm, but with superior stopping power due to heavier bullets and a larger cross-section (.400 vs.356).
Plus S&W didn't lose any investment on the tooling for making 10mm barrels, because now they had a new and improved gun everyone could love -- wonder-9 capacity, in a package most anyone could handle, with stopping power approaching.45 auto.
I'm not sure how much *size* can be changed versus simply "ergonomics".
I just measure my Glock 20 (10mm) and 17 (9mm, same as Beretta 92 load) magazines and grips.
The G20 magazine is 5" in circumference and the grip is a little over 6" in circumference at the midpoint..
The G17 mag is 4" and grip a little over 5".
There's a fair amount of dead space in the base of the backstrap, but the backstrap is angled so the dead space has a lot of taper towards the slide, so adjusting the backstrap is only going to buy you so much benefit because the space for adjustment is reduced as you move up the grip.
I would say there's no meaningful room for reduction in grip width -- the side panels seem to add little width over and above the magazine and it fills the well completely.
Now this may provide the "feel" that makes someone more comfortable even if its not a meaningful reduction in actual size.
They're less useful as clusters grow and workloads migrate across nodes, especially if you use automation like DRS to maintain a self-balancing cluster that levels workloads across clusters.
I've seen them work OK in very small clusters with more or less static node membership, but they're pretty uncommon outside of specialized markets. The last time I've seen one deployed was someplace where they had a crummy, specialized database application that performed poorly and they were tacked on in desperation. The software was the problem and no hardware would fix it.
My fear would be that there is an intelesque hard limit on writes that bricks the drive, and ALL of your drives in a RAID array will hit that limit simultaneously.
All of the drives in that test indicated potential failure via monitoring way before they actually failed, regardless of the actual event of their failure.
My guess is that part of the trick to using an 850 Pro-type drive is close monitoring of drive error status and aggressive replacement of drives showing an indication of failure. It seems unlikely that simultaneous failure would happen across an entire shelf, so closely spaced, as to prevent rebuilds to available hot spares (and hot spare replacement).
The trick is the "seems" part -- AFAIK nobody has publicly done this, created a full shelf array and seen what happens. It would be a great experiment, especially if a lot of data was collected about the workloads the array supported and per-member workloads of the individual disks.
My guess is that replacement rates would be higher than rotational media for ordinary workloads, but less than expected. The question is whether the replacement rate would be offset by the performance benefit. I suspect it might be.
If you could buy a 20 TB all-flash SAN for $50k versus an all-HDD SAN for the same money, would the performance boost be worth it if you had to spend $2000 per year in disk replacement? What's the annual replacement cost (and maybe risk) where it becomes not worth it?
But if you'd rather pay perpetual licensing to third parties for the rest of your natural life (as implied at the end of your post), you're right, I guess having a large amount of local storage doesn't make sense.
It doesn't make a lot of sense, regardless of your FOSS trolling and fantasies. Power, hardware and physical space all cost real money no matter what name is on the login prompt. You live in a fantasy world if you think that managing petabyte scale storage across dozens of independent compute nodes makes any sense at all with FOSS tools off the shelf.
Sure, your specialized team and it's tailored solution could make it work, but I'd love to see the spreadsheet that explains the cost savings of basically going into the SAN business and re-inventing the wheel when you can buy it off the shelf with solutions for problems somebody else has already solved. Unless your team's name is "Google", "Amazon" or maybe "Facebook", few would ever consider "Hey, why buy Compellent or EMC, when we could get a whole bunch of Linux boxes and build our own storage environment. That will surely save money and get us the storage we need faster."
I don't disagree that enterprise storage is costly, which is why hyperconverged compute+storage is becoming a thing (VMware vSAN, MS Storage Spaces, etc), but the reality is that marrying performance and reliability in node-distributed storage is hard, especially if you need immediate coherency between nodes (and not Google-like "fuckit, we're ok with slightly stale pagerank data").
And really, we're talking block storage here -- I don't really know of many IT products more basic and essentially uncustomized than this.
So where are you still seeing monolithic data processing hosts?
As far as I can tell, everyone has moved on to virtualization for all the usual benefits associated with scale out, high availability, disaster recovery. The clients I've run across that need DB crunching IOPS have moved to tiered storage where their DBs live on SSD (like 98%) and a handful which have invested in per-host cache cards, but they're still virtualized and the cache cards were a side result of bad storage/fabric decisions they were stuck with.
I'm sure there are places with the 8U database server because the transaction capacity so very high that no shared storage can accommodate it, but I just don't see it very often.
NVMe only seems to fit outside of the monolithic host if you're drinking the hyperconverged kool-aid and are doing some kind of distributed SAN (like VMware vSAN), where each node supplies part of it. The downside to this, though, always seems to be that the node count ends up being pretty high to get any kind of redundancy for the storage and this ends up being quietly expensive when you start counting node licenses.
I'd love to hear more about how you're using them.
The performance is generally so good that I might be inclined to use a double parity redundancy scheme and hot spare auto rebuild as a hedge against failure. I would generally expect double parity and hot spare rebuild to be fast enough to protect against all but the most catastrophic failures, like all drives somehow failing faster than you can replace spares.
Do tell how you're actually using them and what kind of write usage you se.
I built a tiered storage spaces array and have been mildly alarmed at how great the write rate is on the SSDs I use for the SSD tier, although I think some of that is a function of a shit ton of write churn as I moved a bunch of data around temporarily on that storage space. I need to start keeping a spreadsheet to monitor the TBW over time to see what my estimated lifespan is relative to warranty. If these things are capable of even 1 PB of writes, they should outlast this system's useful lifespan.
Tomorrow: lots of SAS/SATA SSDs with a few PCIe NVMe drives as cache.
Still not seeing the benefits versus complexity and overhead. In a 24 drive shelf you're looking at close to a million read IOPS and sequential reads into the GBytes/second range for sequential reads *just* from SSDs on a SAS backplane.
Maybe there's some exotic, single-host database environment that would benefit from this, but a SSD-only solution would saturate 16GBFC with multipathing. At the point of combining NVMe and SSD, you're now spending more on exotic interconnect fabrics to get the data off the host than you are on the storage.
Given the endurance test people have peformed against "consumer" SSDs, it sure seems like the expected endurance exceeds the warrant by a lot.
This guy:
http://blog.innovaengineering.......has 7 PB written to an 850 Pro and it's still going (last blog update was more than a month ago).
I'd be awful curious to see what the actual durability of an 850 Pro would be in a real production SAN. My suspicion is that the better-than-rated endurance coupled with the low replacement cost might make it worthwhile when you consider the staggering performance you would get.
There might even be gimmicks you could apply on a per-disk basis to improve durability, such as underprovisioning each drive by 25% so that you could wear level across more capacity.
More or less, most storage systems are SAS based and achieve capacity scale and IOPS with many units on a SAS bus.
What's the scaling concept behind this? I'm not aware of a (commonly available) storage expansion system based on PCIe connectivity unless you start getting into something like VSAN or the buzzwordy hyperconverged model where compute nodes create a distributed SAN. But this usually requires a lot of nodes.
This kind of storage seems to aim for single server gross performance, which I guess might be aimed at local caching or for DBs running on a native installed OS in most conventional senses. But if you're in a virtualized environment, this seems to run against the grain somewhat -- DBs utilizing local storage and pinned to nodes with the internal storage or if you're using it as a local cache against a more conventional SAN environment, crippling performance when you move a VM until the new nodes local cache catches up.
I guess I'm not seeing how this is better (other than some gross numbers) than more conventional SAS bus aggregation that achieves IOPS through aggregating individual drives. A dozen conventional 1 TB SSDs will provide similar IOPS, greater aggregate storage and redundancy and with SAS-3 backplane probably even greater throughput.
I kind of liked having a digital library of my DVDs, but when bluerays came around it got much more complicated. I had been using a combination of dvdshrink and/or dvddecoder to rip DVDs, and then handbrake to make m4v, but blueray got complicated and I sort of gave up.
I would argue that the "ala carte" model we're ending up with (at least 6+ streaming services) isn't really ala carte, but more like buffet style. It's all you can eat, but not every buffet serves every item you want, so you have to buy multiple buffets to get a meal.
I'd rather see them come up with per show or per movie pricing, where I pay for every episode or movie I actually watch.
I suspect that even at the inflated Amazon (non-prime) Instant or iTunes pricing, it's getting to the point that unless you have a shedload of time to watch TV, you'd probably be better off not subscribing to a streaming service at all and just buy the content you want as you watch it. At least then you'll only be paying for what you actually watch and not subsidizing (again) the equivalent of 100 channels with nothing on.
Amazon Instant and Netflix seem to have gotten worse in terms of movie selection. I use the hell out of Prime, so I don't care (as much) about it Amazon Instant, but if my kid didn't watch Netflix I would consider dropping it entirely. Plus I seem to remember where Netflix lost/didn't renew a distribution deal with somebody lately, taking away another block of content, leaving even more D grade movies. I occasionally find myself suckered into one them by the description and I'm often baffled how such awful content gets generated with what often amounts to pretty decent production values. It's like they paid for everything but writers, director and actors.
After I wrote the above post I realized I didn't factor in power.
I don't think most SMBs care about power. From my experience, the power savings for them wouldn't even be a consideration. Most might see their (usually inadequate) UPSs gain a few more minutes of run time and their air conditioning work a little better, but by and large it's not really a thing for them.
But at the very large enterprise, I'd almost bet the power savings would be worth the investment in additional floor space if they had to add shelves to meet capacity/density to go flash over HDD.
My sense is, though, that any improvement in flash technology that results in reliability that makes it interchangeable with rotational media will likely be produced in package densities that equal available rotational media. Considering how small existing current flash technology is now, you kind of wonder why you can't get 6 TB SSDs in 3.5" packaging now (besides them costing $3k a pop).
I would say not kaput, but you can probably see it from here.
If innovations like 3D Xpoint from Intel and other "better" flash technologies happen that improve on durability and performance while keeping prices at or below other flash tech, it sure seems to me that the rotational media market will get even thinner.
There may be some market for rotational media as a way of providing vast quantity now in very large tiered storage systems, but these kinds of systems are probably already flash tiered.
But any kind of flash that offers durability closer to rotational media at prices closer to current 3D-NAND type flash disks ought to greatly reduce the middle of the storage market -- SMBs running 1-2 shelves of storage.
Right now they run straight rotational media for the most part because SLC drives in any kind of capacity (native or aggregate) is frightfully expensive. If something like 3D Xpoint lives up to its hype, these kinds of storage systems will be exclusively flash based even if there is a small premium over rotational media. It's hard to say no to a 10-20% price bump when the value add is a half million IOPS and crazy throughput.
I would go out on a limb and wonder what will happen to the price of rotational media when the volume demand drops. Does it mean reduced production capacity and increased prices? About the only customers left for it will be those petabyte scale storage systems where floor space demands density not available in flash drive packaging.
I'm curious if the headline is the exact wording used by apple (unlikely) as it implies there may be an access method that was built in. The encryption wouldn't be broken then.
It probably applies to pre-iOS 8 devices where data protection/whole system encryption wasn't enabled by default but the device is locked by a passcode. IIRC, iOS 8 did more than just enable data protection by default, I think some kinds of changes were made to strengthen the data protection process.
Apple are probably referring not just to whole device encryption but weaknesses in pre-8 encryption processes that allow them to extract decryption keys.
I would think that encryption at the OS level would be a safer concept anyway. It's closer to where the data is actually used and generated and guarantees that the data is encrypted no matter what device a given system is writing to.
It's not hard to see situations where an OS is moved to other hardware or backing storage is changed. Relying on encrypted disks providing that suddenly means it's unencrypted.
Slashdot Mirror
Could they use the heated containment water to drive a Stirling engine which would pump more water?
I'm not a nuclear scientist (let alone a rocket scientist), but somehow it seems like there's a way to (relatively simply) use the heat from containment cooling to pump water.
And is there a reason the containment water couldn't be a loop with a cooling stage so it could be self-replenishing? It seems to make more sense if you consider the idea of the containment heat being used to drive pumps which circulated the water.
I'm sure there are good reasons why this is a Rube Goldberg perpetual motion kind of an idea, but it seems like a way where as long as it generates heat it could pump its own cooling.
It would seem to be pretty trivial to implement a feature where the phone is aware of the account's data plan details (used, available, etc).
The phone could then have some user configurable warning threshold when the plan's available data goes below that threshold.
Apps and features (like wifi-assist) could then have a setting that allow/prohibits them to use data when the threshold has been exceeded.
If there was some standardized method of obtaining and communicating this info, shared data plans could manage these settings on the carrier account page per device -- individual caps, thresholds relative to individual caps or total shared caps.
This sounds like an application of a de minimis principal -- arguing that the scale of Wikimedia's traffic isn't large enough to give it standing.
I seem to remember learning this phrase in college in a political science class about some guy who sued the government to get some kind of information on defense appropriations, but was rejected because his individual contribution was so small that it essentially didn't matter.
I sometimes wonder if the answer is something we don't want it to be -- that heavy handed policing is repugnant, but the nature of our society is that the poor, urban populations suffer from so much social malaise that without heavy handed policing they will tear themselves apart while inflicting collateral damage on the rest of society through crime and violence?
The causes of malaise are often unjust -- discrimination, lack of opportunity, but also include self-inflicted problems of unplanned/bad parenting, purposeful rejection of positive social choices (dropping out of school, etc) and so on. It's not completely their fault, but it's not completely a question of inescapable victimhood, either.
Heavy-handed policing is likely not the best course to *solve* these problems, but the scale and nature of them is such that the costs and scale of the social welfare solutions which could possibly be more effective are seen as an impossible burden (which itself is a byproduct of economic inequality and ineffective governance).
Heavy-handed policing is thus seen as the most obtainable possible solution -- the least worst possible alternative.
In other words, we may hate heavy handed policing because it also as collateral damage of all kinds, including gross injustices, but without it we may have a kind of chaos that ends up being worse.
All of this just seems to reveal the genius that was expressed in John Browning's .45 auto 1911. Grip size, grip angle, cartridge effectiveness, felt recoil, accuracy -- it's all so much greater than the sum of it parts.
And despite the "advancements" of other rounds and gun designs, I have yet to shoot an automatic that doesn't feel like something got sacrificed for whatever advancement a specific gun brought to the table. All double-stacks seem to compromise ergonomics in some way through larger grips, and most of the "other" cartridges try to compensate for smaller bullet cross section and reduced case capacity with high-pressure loads in an attempt to throw their smaller, lighter projectiles fast enough to get sufficient energy yet rely on highly engineered projectiles to prevent over-penetration.
I have large enough hands and enough body mass that some of these advancements are less of a trade-off for me personally (I have carried a Glock 29 for years), but it's still patently obvious just how wonderful the 1911 is.
The only thing I'm not a huge fan of is carrying a 1911 cocked and locked. I believe the design, but personal hesitancy makes it a non-starter for me. It doesn't help that when I was first introduced to it by a more experienced shooter and questioned the safety that the gun being demonstrated actually managed to drop the hammer without the grip safety (a problem with that specific gun's parts). The gun was unloaded and there were no consequence, but you might equate it to bad experience with a dog when you're young.
If you're not searching for a specific brand name of something, you will often get 87 nearly identical products (often with wildly varying ratings, chock full o'astroturfed reviews) for junk made in China.
Half the time I search on Amazon I feel like I'm just getting an iframe with the results from Alibaba.
Forgetting the politics, economics and global events surrounding those $25 trillion, if we *had* decided to spend $25 trillion (which is nearly 2 years of total US GDP) on fusion would we have even a small scale utility version by now?
I'm assuming that given the size of that amount of money and the actual US GDP, it would take a massive effort just to practically even spend that kind of money, let alone perform the science and engineering it might enable (such as building facilities, etc).
The Samsung stick has only 20% of the speed of DDR3 and enough extra latency that it would be a painful step backward.
Although isn't that basically the system used by smartphones and tablets? Suspend the app and page it to flash when you switch to another context? But they also have much smaller applications and have like 1-2 GM RAM to begin with.
I think where that system might work better overall is on virtualization hosts where active memory is much lower than assigned memory. Those inactive pages available on a fast swap drive would make high density hosts a lot less painful than generic disk swapping.
The FBI had standardized on .38 special and got their asses kicked in the infamous "Miami shootout".
The FBI turned to S&W who worked with Norma to develop the 10mm round which got put in the S&W 10xx series of autoloaders, sharing the same frame as the 45xx series of .45s.
The full-power Norma 10mm was hella stopping power, but it was unpopular with most women and some men due to size and recoil. The original Norma loads were close to .41 Magnum power. I handload both cartridges and believe this is pretty true -- my 610 with maxed-out 10mms feel about the same as my Model 57 with middling loads.
The FBI decided they didn't like the 10mm afterall, but some bright bulb at S&W realized that if you cut down the 10mm case by a few mm you could have a grip size and recoil similar to 9mm, but with superior stopping power due to heavier bullets and a larger cross-section (.400 vs .356).
Plus S&W didn't lose any investment on the tooling for making 10mm barrels, because now they had a new and improved gun everyone could love -- wonder-9 capacity, in a package most anyone could handle, with stopping power approaching .45 auto.
And thus .40S&W was born.
I'm not sure how much *size* can be changed versus simply "ergonomics".
I just measure my Glock 20 (10mm) and 17 (9mm, same as Beretta 92 load) magazines and grips.
The G20 magazine is 5" in circumference and the grip is a little over 6" in circumference at the midpoint..
The G17 mag is 4" and grip a little over 5".
There's a fair amount of dead space in the base of the backstrap, but the backstrap is angled so the dead space has a lot of taper towards the slide, so adjusting the backstrap is only going to buy you so much benefit because the space for adjustment is reduced as you move up the grip.
I would say there's no meaningful room for reduction in grip width -- the side panels seem to add little width over and above the magazine and it fills the well completely.
Now this may provide the "feel" that makes someone more comfortable even if its not a meaningful reduction in actual size.
They're less useful as clusters grow and workloads migrate across nodes, especially if you use automation like DRS to maintain a self-balancing cluster that levels workloads across clusters.
I've seen them work OK in very small clusters with more or less static node membership, but they're pretty uncommon outside of specialized markets. The last time I've seen one deployed was someplace where they had a crummy, specialized database application that performed poorly and they were tacked on in desperation. The software was the problem and no hardware would fix it.
My fear would be that there is an intelesque hard limit on writes that bricks the drive, and ALL of your drives in a RAID array will hit that limit simultaneously.
All of the drives in that test indicated potential failure via monitoring way before they actually failed, regardless of the actual event of their failure.
My guess is that part of the trick to using an 850 Pro-type drive is close monitoring of drive error status and aggressive replacement of drives showing an indication of failure. It seems unlikely that simultaneous failure would happen across an entire shelf, so closely spaced, as to prevent rebuilds to available hot spares (and hot spare replacement).
The trick is the "seems" part -- AFAIK nobody has publicly done this, created a full shelf array and seen what happens. It would be a great experiment, especially if a lot of data was collected about the workloads the array supported and per-member workloads of the individual disks.
My guess is that replacement rates would be higher than rotational media for ordinary workloads, but less than expected. The question is whether the replacement rate would be offset by the performance benefit. I suspect it might be.
If you could buy a 20 TB all-flash SAN for $50k versus an all-HDD SAN for the same money, would the performance boost be worth it if you had to spend $2000 per year in disk replacement? What's the annual replacement cost (and maybe risk) where it becomes not worth it?
But if you'd rather pay perpetual licensing to third parties for the rest of your natural life (as implied at the end of your post), you're right, I guess having a large amount of local storage doesn't make sense.
It doesn't make a lot of sense, regardless of your FOSS trolling and fantasies. Power, hardware and physical space all cost real money no matter what name is on the login prompt. You live in a fantasy world if you think that managing petabyte scale storage across dozens of independent compute nodes makes any sense at all with FOSS tools off the shelf.
Sure, your specialized team and it's tailored solution could make it work, but I'd love to see the spreadsheet that explains the cost savings of basically going into the SAN business and re-inventing the wheel when you can buy it off the shelf with solutions for problems somebody else has already solved. Unless your team's name is "Google", "Amazon" or maybe "Facebook", few would ever consider "Hey, why buy Compellent or EMC, when we could get a whole bunch of Linux boxes and build our own storage environment. That will surely save money and get us the storage we need faster."
I don't disagree that enterprise storage is costly, which is why hyperconverged compute+storage is becoming a thing (VMware vSAN, MS Storage Spaces, etc), but the reality is that marrying performance and reliability in node-distributed storage is hard, especially if you need immediate coherency between nodes (and not Google-like "fuckit, we're ok with slightly stale pagerank data").
And really, we're talking block storage here -- I don't really know of many IT products more basic and essentially uncustomized than this.
So where are you still seeing monolithic data processing hosts?
As far as I can tell, everyone has moved on to virtualization for all the usual benefits associated with scale out, high availability, disaster recovery. The clients I've run across that need DB crunching IOPS have moved to tiered storage where their DBs live on SSD (like 98%) and a handful which have invested in per-host cache cards, but they're still virtualized and the cache cards were a side result of bad storage/fabric decisions they were stuck with.
I'm sure there are places with the 8U database server because the transaction capacity so very high that no shared storage can accommodate it, but I just don't see it very often.
NVMe only seems to fit outside of the monolithic host if you're drinking the hyperconverged kool-aid and are doing some kind of distributed SAN (like VMware vSAN), where each node supplies part of it. The downside to this, though, always seems to be that the node count ends up being pretty high to get any kind of redundancy for the storage and this ends up being quietly expensive when you start counting node licenses.
I'd love to hear more about how you're using them.
The performance is generally so good that I might be inclined to use a double parity redundancy scheme and hot spare auto rebuild as a hedge against failure. I would generally expect double parity and hot spare rebuild to be fast enough to protect against all but the most catastrophic failures, like all drives somehow failing faster than you can replace spares.
Do tell how you're actually using them and what kind of write usage you se.
I built a tiered storage spaces array and have been mildly alarmed at how great the write rate is on the SSDs I use for the SSD tier, although I think some of that is a function of a shit ton of write churn as I moved a bunch of data around temporarily on that storage space. I need to start keeping a spreadsheet to monitor the TBW over time to see what my estimated lifespan is relative to warranty. If these things are capable of even 1 PB of writes, they should outlast this system's useful lifespan.
Tomorrow: lots of SAS/SATA SSDs with a few PCIe NVMe drives as cache.
Still not seeing the benefits versus complexity and overhead. In a 24 drive shelf you're looking at close to a million read IOPS and sequential reads into the GBytes/second range for sequential reads *just* from SSDs on a SAS backplane.
Maybe there's some exotic, single-host database environment that would benefit from this, but a SSD-only solution would saturate 16GBFC with multipathing. At the point of combining NVMe and SSD, you're now spending more on exotic interconnect fabrics to get the data off the host than you are on the storage.
Given the endurance test people have peformed against "consumer" SSDs, it sure seems like the expected endurance exceeds the warrant by a lot.
This guy:
http://blog.innovaengineering.... ...has 7 PB written to an 850 Pro and it's still going (last blog update was more than a month ago).
I'd be awful curious to see what the actual durability of an 850 Pro would be in a real production SAN. My suspicion is that the better-than-rated endurance coupled with the low replacement cost might make it worthwhile when you consider the staggering performance you would get.
There might even be gimmicks you could apply on a per-disk basis to improve durability, such as underprovisioning each drive by 25% so that you could wear level across more capacity.
More or less, most storage systems are SAS based and achieve capacity scale and IOPS with many units on a SAS bus.
What's the scaling concept behind this? I'm not aware of a (commonly available) storage expansion system based on PCIe connectivity unless you start getting into something like VSAN or the buzzwordy hyperconverged model where compute nodes create a distributed SAN. But this usually requires a lot of nodes.
This kind of storage seems to aim for single server gross performance, which I guess might be aimed at local caching or for DBs running on a native installed OS in most conventional senses. But if you're in a virtualized environment, this seems to run against the grain somewhat -- DBs utilizing local storage and pinned to nodes with the internal storage or if you're using it as a local cache against a more conventional SAN environment, crippling performance when you move a VM until the new nodes local cache catches up.
I guess I'm not seeing how this is better (other than some gross numbers) than more conventional SAS bus aggregation that achieves IOPS through aggregating individual drives. A dozen conventional 1 TB SSDs will provide similar IOPS, greater aggregate storage and redundancy and with SAS-3 backplane probably even greater throughput.
Eduncate me, please.
Can handbrake rip copy protectet blurays?
I kind of liked having a digital library of my DVDs, but when bluerays came around it got much more complicated. I had been using a combination of dvdshrink and/or dvddecoder to rip DVDs, and then handbrake to make m4v, but blueray got complicated and I sort of gave up.
I would argue that the "ala carte" model we're ending up with (at least 6+ streaming services) isn't really ala carte, but more like buffet style. It's all you can eat, but not every buffet serves every item you want, so you have to buy multiple buffets to get a meal.
I'd rather see them come up with per show or per movie pricing, where I pay for every episode or movie I actually watch.
I suspect that even at the inflated Amazon (non-prime) Instant or iTunes pricing, it's getting to the point that unless you have a shedload of time to watch TV, you'd probably be better off not subscribing to a streaming service at all and just buy the content you want as you watch it. At least then you'll only be paying for what you actually watch and not subsidizing (again) the equivalent of 100 channels with nothing on.
Amazon Instant and Netflix seem to have gotten worse in terms of movie selection. I use the hell out of Prime, so I don't care (as much) about it Amazon Instant, but if my kid didn't watch Netflix I would consider dropping it entirely. Plus I seem to remember where Netflix lost/didn't renew a distribution deal with somebody lately, taking away another block of content, leaving even more D grade movies. I occasionally find myself suckered into one them by the description and I'm often baffled how such awful content gets generated with what often amounts to pretty decent production values. It's like they paid for everything but writers, director and actors.
After I wrote the above post I realized I didn't factor in power.
I don't think most SMBs care about power. From my experience, the power savings for them wouldn't even be a consideration. Most might see their (usually inadequate) UPSs gain a few more minutes of run time and their air conditioning work a little better, but by and large it's not really a thing for them.
But at the very large enterprise, I'd almost bet the power savings would be worth the investment in additional floor space if they had to add shelves to meet capacity/density to go flash over HDD.
My sense is, though, that any improvement in flash technology that results in reliability that makes it interchangeable with rotational media will likely be produced in package densities that equal available rotational media. Considering how small existing current flash technology is now, you kind of wonder why you can't get 6 TB SSDs in 3.5" packaging now (besides them costing $3k a pop).
I would say not kaput, but you can probably see it from here.
If innovations like 3D Xpoint from Intel and other "better" flash technologies happen that improve on durability and performance while keeping prices at or below other flash tech, it sure seems to me that the rotational media market will get even thinner.
There may be some market for rotational media as a way of providing vast quantity now in very large tiered storage systems, but these kinds of systems are probably already flash tiered.
But any kind of flash that offers durability closer to rotational media at prices closer to current 3D-NAND type flash disks ought to greatly reduce the middle of the storage market -- SMBs running 1-2 shelves of storage.
Right now they run straight rotational media for the most part because SLC drives in any kind of capacity (native or aggregate) is frightfully expensive. If something like 3D Xpoint lives up to its hype, these kinds of storage systems will be exclusively flash based even if there is a small premium over rotational media. It's hard to say no to a 10-20% price bump when the value add is a half million IOPS and crazy throughput.
I would go out on a limb and wonder what will happen to the price of rotational media when the volume demand drops. Does it mean reduced production capacity and increased prices? About the only customers left for it will be those petabyte scale storage systems where floor space demands density not available in flash drive packaging.
It's like the UK is reliving it's past, minus the part about the Magna Carta.
I'm curious if the headline is the exact wording used by apple (unlikely) as it implies there may be an access method that was built in. The encryption wouldn't be broken then.
It probably applies to pre-iOS 8 devices where data protection/whole system encryption wasn't enabled by default but the device is locked by a passcode. IIRC, iOS 8 did more than just enable data protection by default, I think some kinds of changes were made to strengthen the data protection process.
Apple are probably referring not just to whole device encryption but weaknesses in pre-8 encryption processes that allow them to extract decryption keys.
I would think that encryption at the OS level would be a safer concept anyway. It's closer to where the data is actually used and generated and guarantees that the data is encrypted no matter what device a given system is writing to.
It's not hard to see situations where an OS is moved to other hardware or backing storage is changed. Relying on encrypted disks providing that suddenly means it's unencrypted.