The difference is that most of those people generally worked in a capitalist system (Roman empire) in order to get the money and goods they donated.
Communal type thinking can sort of work in those situations where everyone shares similar values, and it is a small enough community to deal with folks who arent sincere and simply want to exploit the system, and where it is voluntary. Bigger / national scale, not so much.
In the same way that capitalism has been tried and failed because you can't remove greed from the human condition.
Are you sure we are talking about the same capitalism? The capitalism I know uses greed / necessity (or a desire to make more for philanthropic reasons) as the bait to get people to do things.
It would WORK without greed, but it hardly is hampered by it, and at least it attempts to deal with greed rather than pretending it doesnt exist.
Someone will always take advantage of their position and grab more than they deserve leaving the less able or unfortunate with less than they deserve.
I think the number of people with a lot of money who "dont deserve it" is far less than is imagined. What, you think those people took no risks and took some easy road to money? Funny, you would think a lot more people would do that if it were so simple.
... And that every attempt to do better than capitalism has always resulted in worse evils than capitalism brings, yes.
Or do you have a better motivator than wealth for getting people to be productive? The very fact that you realize wealth trumps all else (and accuse that as being the cause of capitalism), is itself THE argument for capitalism.
It is very hard to deny the motivational power of wealth and power; capitalism simply accepts this and tries to harness it for mutual societal good, rather than most other systems which pretend that noone is spurred on by power, and end up as authoritarian nightmares.
That really doesnt work so well for sites like Google / all their services, or Amazon, that people may want to access from various places and on various computers. Are you suggesting that we teach everyone about the concepts of certificates, thumbprints, and trust, so that they can pore over the certificate trust chains on each computer they ever want to use?
Not when there is a very clear answer in the documents that grant our government its authority. Its not a far step from "we can ignore the 9th and 10th amendments" to "what 1st and 2nd amendments?" If this was something we really wanted to change, an amendment would be the proper way, expanding what "general welfare" means.
When lasers hit imperfectly reflective surfaces, they scatter high intensity light.
The point everyone is trying to make to you is that when you shine that laser at a brick or concrete wall (which has shiney, irregular bits in it), and 10 people look at that dot, those 10 people are all now blind. If the light glances into someones eye from an angle, they now have eye damage.
And even if you HAD to aim the light at someone's eye, the laser is continuous, and the trigger can be held for at least a second or two, and swept in a direction with a casual wave of the hand. Anyone along that path can have serious eye damage done.
The limiting factor may be space. If the 1x1TB drive fits into a 1U enclosure but the 2x600GB drive requires a 2U enclosure, it may not be worth doubling your datacenter rack space for the storage.
If your only two comparisons here are between 15k drives and SSDs, and space is the limiter, your scope is too limited. Why not do 7200RPM drives, which go up to 3TB per drive? Theyre slower, but substantially cheaper per GB, and if reliability is an issue you can get about 2TB of RAID1 space on "enterprise" drives for about $4-600, in a very small amount of space.
If the limiting factor is reliability, I would REALLY caution against hopping on the SSD ship. Jury is still out on that, and initial reports seem to be "odd random failures of the entire drive after a year or so". Not sure if they are isolated, but looks at AnandTech, Newegg reviews, etc leave a scary impression. On the other hand, I just yesterday worked on a Wang Mainframe with a 27 year old mechanical drive that JUST failed. I also regularly deal with 5+ year old mechanical drives in servers, that work fine-- failures are the exception, not the norm. Id say thats a darn good track record, and data recovery off of failed drives is a known quantity at this point. SSDs require, if memory serves, scanning electron microscopes in the worst case scenario.
Even comparing two RAID0 arrays, SSD should be more reliable.
I will want to see a lot of use cases, whitepapers, and real world data before I even think of suggesting that in a client desktop, let alone a server. That seems to be the definition of playing with fire-- double your failure rate on an unrecoverable medium, with an unproven track record. No thanks.
Wait, so now we need TWO of those SSDs in a RAID1? But now the cost is twice as high, and write speed hasnt increased; meanwhile, we've just scaled the SAS drive array up to 16 spindles.
Im also curious, do they make hot-swap cages for PCIe SSD cards? Or do we need to power the whole array down in order to swap out this beastly card when it fails next year?
That is correct. A lot of SSDs use RAID0 internally, which means for every RAID block of data that goes bad on one SSD, double (or quadruple, depending on the size of the array) that amount of data is lost. It starts getting pretty scary pretty fast. RAID1 the entire thing and you really have an odd sort of RAID10 going on, but without the assurance that its a "proper" RAID 10.
Also, I will note that 450GB SAS drives start around $300, which is even cheaper. At that point you can get 3.2TB of data for the price of that SSD, and then RAID 10 the thing, and you now have hotswappability, very good performance, and still 50% more storage than the SSD solution. You lose a bit on the seek performance, but there is no way in the world that SSD is going to beat 8x 15k SAS drives in raw throughput.
You also dont have to worry about random failure in about 1.5 years because noone knows what the real-world MTBF is on these SSDs, but noone seems to want to think about that little issue.
First, I dont think SSD latencies are generally measured in ms-- theyre supposed to be much, much faster than that. If the comparison is "grab this random sector asap", there is no situation I can think of where the mechanical drive will come within even a 50% margin of beating the SSD.
But the point is that for larger chunks, or easily queuable chunks, the mechanical drive can have much better performance than its latency seems to suggest-- a lot better than simply "1000 / [seek latency]" ops per second.
Depends heavily on which two drives you are comparing. unTRIM'd usually kills performance pretty badly.
For instance, check this 2009 article, back before TRIM was in standard use. Note how the OCZ Summit drive drops from 12MB/s in their test to about 2.5MB/s-- about 1/4 the speed.
Considering that SATA drives with NCQ can push over 150MB/s, and most SSDs arent quite up to 600MB/s yet, there certainly are places where lack of TRIM will cause the mechanical drive to be better-- especially since, if the drive doesnt support TRIM or any garbage collection, its speed will eventually drop permanently below the speed of the mechanical.
Because I think the governments role and the lines that need to be drawn should be done so on the basis of "is it necessary or urgent, and is it something the government is uniquely able to do".
With TV, I think the answers are clearly "No, no, and No".
You have to aim that gun, however, and as has been pointed out can feasibly injure or kill much fewer people in a given period of time with the gun; additionally, with the gun you basically need rather good aim and an intention to harm (or youre foolish enough to play with a loaded gun with the safety off in public). With the laser, even "normal" use outside can instantly (speed of light) and silently (no clue as to what is happening unless you see the laser, and then, whoops, youre blind) injure a large number of people.
I think you will also find that the laser's ability to injure diminishes FAR less with distance than the handgun.
That 100 seeks per second isnt the entire story, however-- basically every drive you can possibly get these days supports Native Command Queuing, which means that the drive will try to organize its reads so that it doesnt have to seek to position 50 to grab a block of data, then reseek to position 40 to grab another block. With NCQ, it would rearrange the requests so that it first gets position 40, then goes to 50 within the same "rotation"-- so both requests were done in under that 10ms.
NCQ makes a huge difference, and goes a long way to changing the real story. (also, apparently its called TCQ on SAS / SCSI devices).
Whats the price difference between 2 600GB mechanicals and the 1 1TB drive?
Also, whats MTBF, and is that SSD using internal RAID0 like I think it is? If so, have fun with data recovery when one flash failure kills the entire 1TB.
At $30,000 per SSD, times 3, you get $90,000. Divide that by 60 and you get $1,500 per hard drive to break even.
Im not sure Ive ever seen a hard drive that costs $1,500; Newegg says 450GB SAS 15k drives can be had for $300.
But then, we dont know how much data is being stored by that SSD, or how much was being stored by the mechanicals, or how much parity (if any?) each system had, or whether they were from the same vendor... all of which make the article pretty darn useless.
They mention, for example, that their SAN totals 900TB. Im almost positive that at $30k a pop those SSDs are NOT holding 300TB each (and if they are, someone point me to the vendor fast, so I can quit my job and start a company serving up high-speed database storage).
It depends on the application, and how it stores its data. Not all scenarios are conducive to rsync; it is possible for an rsync to take more time than just doing a copy.
I've lusted after 1000mW diodes for a long time with the idea of programming beam splitters, mirrors and the like to produce laser shows for parties and family events but it's just too dangerous for close quarters. Kids will insist on playing with them, or if I goof, or anything fails, the idea of blinding someone is scary.
Arent laser-light shows-- especially homemade ones-- also spectacularly dangerous?
A gun is pretty limited by the noise it makes, and by the fact that you have to aim it. Aim the laser at a prism (or basically anything semi-reflective-- white paint on the side of a house?) and watch as everyone around you gets eye damage, silently, immediately. Guns also have a very limited range at which they are effective-- handguns are well under a kilometer i believe. Im sure this laser is effective at well over a mile.
And unlike gunshot wounds, im pretty sure eye damage from lasers is generally permenant.
I think if you dont have parents that want to raise their kids and would rather have the states do it, you have a problem.
Part of the issue is that a lot of parents see the goal in life as making as much money as possible while their kids are in school; how many sit down and try to budget and see if one of the parents can stay at home (ohnoes, a stay at home mom!), and raise the kids, and teach them? For many (certainly this was my experience, never seeing homeschoolers until a few years ago), it never was part of the equation.
So no, I dont have an answer as to how to extend it to the population, as you cant legislate parents wanting to raise their kids. But for the parents in a situation where they simply have to work (single parents etc), there are a few options-- vouchers, private schools, and possibly legislation that gives tax credits to the parents for homeschooling (perhaps giving them 75% of whatever the normal cost is for the state to educate the child-- win win). Such tax credit legislation may already exist, by the way, im not familiar with all of the homeschool laws and what not.
As for homeschooling only being an option for the wealthy, one of my siblings is / was homeschooling, and you wouldnt really call them wealthy. They simply budget for what is really important-- their children's upbringing and education, rather than pursuit of wealth. This is not an isolated incident, I know a number of people who have to budget for it, but in the end it appears to be worth it, as the kids tend to be incredibly well-adjusted, intelligent, and respectful, and have wonderful relationships with their parents as they actually get to see them more than 4 hours a day.
Slashdot Mirror
The difference is that most of those people generally worked in a capitalist system (Roman empire) in order to get the money and goods they donated.
Communal type thinking can sort of work in those situations where everyone shares similar values, and it is a small enough community to deal with folks who arent sincere and simply want to exploit the system, and where it is voluntary. Bigger / national scale, not so much.
In the same way that capitalism has been tried and failed because you can't remove greed from the human condition.
Are you sure we are talking about the same capitalism? The capitalism I know uses greed / necessity (or a desire to make more for philanthropic reasons) as the bait to get people to do things.
It would WORK without greed, but it hardly is hampered by it, and at least it attempts to deal with greed rather than pretending it doesnt exist.
Someone will always take advantage of their position and grab more than they deserve leaving the less able or unfortunate with less than they deserve.
I think the number of people with a lot of money who "dont deserve it" is far less than is imagined. What, you think those people took no risks and took some easy road to money? Funny, you would think a lot more people would do that if it were so simple.
... And that every attempt to do better than capitalism has always resulted in worse evils than capitalism brings, yes.
Or do you have a better motivator than wealth for getting people to be productive? The very fact that you realize wealth trumps all else (and accuse that as being the cause of capitalism), is itself THE argument for capitalism.
It is very hard to deny the motivational power of wealth and power; capitalism simply accepts this and tries to harness it for mutual societal good, rather than most other systems which pretend that noone is spurred on by power, and end up as authoritarian nightmares.
That really doesnt work so well for sites like Google / all their services, or Amazon, that people may want to access from various places and on various computers. Are you suggesting that we teach everyone about the concepts of certificates, thumbprints, and trust, so that they can pore over the certificate trust chains on each computer they ever want to use?
Not when there is a very clear answer in the documents that grant our government its authority. Its not a far step from "we can ignore the 9th and 10th amendments" to "what 1st and 2nd amendments?" If this was something we really wanted to change, an amendment would be the proper way, expanding what "general welfare" means.
When lasers hit imperfectly reflective surfaces, they scatter high intensity light.
The point everyone is trying to make to you is that when you shine that laser at a brick or concrete wall (which has shiney, irregular bits in it), and 10 people look at that dot, those 10 people are all now blind. If the light glances into someones eye from an angle, they now have eye damage.
And even if you HAD to aim the light at someone's eye, the laser is continuous, and the trigger can be held for at least a second or two, and swept in a direction with a casual wave of the hand. Anyone along that path can have serious eye damage done.
The limiting factor may be space. If the 1x1TB drive fits into a 1U enclosure but the 2x600GB drive requires a 2U enclosure, it may not be worth doubling your datacenter rack space for the storage.
If your only two comparisons here are between 15k drives and SSDs, and space is the limiter, your scope is too limited. Why not do 7200RPM drives, which go up to 3TB per drive? Theyre slower, but substantially cheaper per GB, and if reliability is an issue you can get about 2TB of RAID1 space on "enterprise" drives for about $4-600, in a very small amount of space.
If the limiting factor is reliability, I would REALLY caution against hopping on the SSD ship. Jury is still out on that, and initial reports seem to be "odd random failures of the entire drive after a year or so". Not sure if they are isolated, but looks at AnandTech, Newegg reviews, etc leave a scary impression. On the other hand, I just yesterday worked on a Wang Mainframe with a 27 year old mechanical drive that JUST failed. I also regularly deal with 5+ year old mechanical drives in servers, that work fine-- failures are the exception, not the norm. Id say thats a darn good track record, and data recovery off of failed drives is a known quantity at this point. SSDs require, if memory serves, scanning electron microscopes in the worst case scenario.
Even comparing two RAID0 arrays, SSD should be more reliable.
I will want to see a lot of use cases, whitepapers, and real world data before I even think of suggesting that in a client desktop, let alone a server. That seems to be the definition of playing with fire-- double your failure rate on an unrecoverable medium, with an unproven track record. No thanks.
Wait, so now we need TWO of those SSDs in a RAID1? But now the cost is twice as high, and write speed hasnt increased; meanwhile, we've just scaled the SAS drive array up to 16 spindles.
Im also curious, do they make hot-swap cages for PCIe SSD cards? Or do we need to power the whole array down in order to swap out this beastly card when it fails next year?
That is correct. A lot of SSDs use RAID0 internally, which means for every RAID block of data that goes bad on one SSD, double (or quadruple, depending on the size of the array) that amount of data is lost. It starts getting pretty scary pretty fast. RAID1 the entire thing and you really have an odd sort of RAID10 going on, but without the assurance that its a "proper" RAID 10.
Also, I will note that 450GB SAS drives start around $300, which is even cheaper. At that point you can get 3.2TB of data for the price of that SSD, and then RAID 10 the thing, and you now have hotswappability, very good performance, and still 50% more storage than the SSD solution. You lose a bit on the seek performance, but there is no way in the world that SSD is going to beat 8x 15k SAS drives in raw throughput.
You also dont have to worry about random failure in about 1.5 years because noone knows what the real-world MTBF is on these SSDs, but noone seems to want to think about that little issue.
Thats true, and a good point; but the story does mention that they have 900TB of data to store. Im pretty sure its not all sitting on SSDs.
First, I dont think SSD latencies are generally measured in ms-- theyre supposed to be much, much faster than that. If the comparison is "grab this random sector asap", there is no situation I can think of where the mechanical drive will come within even a 50% margin of beating the SSD.
But the point is that for larger chunks, or easily queuable chunks, the mechanical drive can have much better performance than its latency seems to suggest-- a lot better than simply "1000 / [seek latency]" ops per second.
Depends heavily on which two drives you are comparing. unTRIM'd usually kills performance pretty badly.
For instance, check this 2009 article, back before TRIM was in standard use. Note how the OCZ Summit drive drops from 12MB/s in their test to about 2.5MB/s-- about 1/4 the speed.
Considering that SATA drives with NCQ can push over 150MB/s, and most SSDs arent quite up to 600MB/s yet, there certainly are places where lack of TRIM will cause the mechanical drive to be better-- especially since, if the drive doesnt support TRIM or any garbage collection, its speed will eventually drop permanently below the speed of the mechanical.
Because I think the governments role and the lines that need to be drawn should be done so on the basis of "is it necessary or urgent, and is it something the government is uniquely able to do".
With TV, I think the answers are clearly "No, no, and No".
You have to aim that gun, however, and as has been pointed out can feasibly injure or kill much fewer people in a given period of time with the gun; additionally, with the gun you basically need rather good aim and an intention to harm (or youre foolish enough to play with a loaded gun with the safety off in public). With the laser, even "normal" use outside can instantly (speed of light) and silently (no clue as to what is happening unless you see the laser, and then, whoops, youre blind) injure a large number of people.
I think you will also find that the laser's ability to injure diminishes FAR less with distance than the handgun.
Sure there is, if your drive doesnt support TRIM (or hasnt TRIM'd in a while)-- that can certainly result in a slower drive.
That 100 seeks per second isnt the entire story, however-- basically every drive you can possibly get these days supports Native Command Queuing, which means that the drive will try to organize its reads so that it doesnt have to seek to position 50 to grab a block of data, then reseek to position 40 to grab another block. With NCQ, it would rearrange the requests so that it first gets position 40, then goes to 50 within the same "rotation"-- so both requests were done in under that 10ms.
NCQ makes a huge difference, and goes a long way to changing the real story. (also, apparently its called TCQ on SAS / SCSI devices).
The tester was willing to test only 3 SSD's versus *60* 15K drives.
Except they COMPLETELY neglected to mention whether there was parity in place on either system, or how much data was being held by each system.
Protip-- the mechanicals were holding a TON more data than the SSD systems, and im pretty sure the mechanicals can be hotswapped.
The only mildly surprising part is that part about the HDDs winning for all writes,
Maybe their SSD doesnt support TRIM. That would certainly explain it.
Whats the price difference between 2 600GB mechanicals and the 1 1TB drive?
Also, whats MTBF, and is that SSD using internal RAID0 like I think it is? If so, have fun with data recovery when one flash failure kills the entire 1TB.
At $30,000 per SSD, times 3, you get $90,000. Divide that by 60 and you get $1,500 per hard drive to break even.
Im not sure Ive ever seen a hard drive that costs $1,500; Newegg says 450GB SAS 15k drives can be had for $300.
But then, we dont know how much data is being stored by that SSD, or how much was being stored by the mechanicals, or how much parity (if any?) each system had, or whether they were from the same vendor... all of which make the article pretty darn useless.
They mention, for example, that their SAN totals 900TB. Im almost positive that at $30k a pop those SSDs are NOT holding 300TB each (and if they are, someone point me to the vendor fast, so I can quit my job and start a company serving up high-speed database storage).
It depends on the application, and how it stores its data. Not all scenarios are conducive to rsync; it is possible for an rsync to take more time than just doing a copy.
I've lusted after 1000mW diodes for a long time with the idea of programming beam splitters, mirrors and the like to produce laser shows for parties and family events but it's just too dangerous for close quarters. Kids will insist on playing with them, or if I goof, or anything fails, the idea of blinding someone is scary.
Arent laser-light shows-- especially homemade ones-- also spectacularly dangerous?
Its significantly harder and messier to kill someone with a hammer than it is to blind someone with a laser.
A gun is pretty limited by the noise it makes, and by the fact that you have to aim it. Aim the laser at a prism (or basically anything semi-reflective-- white paint on the side of a house?) and watch as everyone around you gets eye damage, silently, immediately. Guns also have a very limited range at which they are effective-- handguns are well under a kilometer i believe. Im sure this laser is effective at well over a mile.
And unlike gunshot wounds, im pretty sure eye damage from lasers is generally permenant.
I think if you dont have parents that want to raise their kids and would rather have the states do it, you have a problem.
Part of the issue is that a lot of parents see the goal in life as making as much money as possible while their kids are in school; how many sit down and try to budget and see if one of the parents can stay at home (ohnoes, a stay at home mom!), and raise the kids, and teach them? For many (certainly this was my experience, never seeing homeschoolers until a few years ago), it never was part of the equation.
So no, I dont have an answer as to how to extend it to the population, as you cant legislate parents wanting to raise their kids. But for the parents in a situation where they simply have to work (single parents etc), there are a few options-- vouchers, private schools, and possibly legislation that gives tax credits to the parents for homeschooling (perhaps giving them 75% of whatever the normal cost is for the state to educate the child-- win win). Such tax credit legislation may already exist, by the way, im not familiar with all of the homeschool laws and what not.
As for homeschooling only being an option for the wealthy, one of my siblings is / was homeschooling, and you wouldnt really call them wealthy. They simply budget for what is really important-- their children's upbringing and education, rather than pursuit of wealth. This is not an isolated incident, I know a number of people who have to budget for it, but in the end it appears to be worth it, as the kids tend to be incredibly well-adjusted, intelligent, and respectful, and have wonderful relationships with their parents as they actually get to see them more than 4 hours a day.