I guess you're unfamiliar with Bezos and unfamiliar with how and why Amazon began. Books are not what makes Amazon special. The idea of books came after Bezos designed the system that makes Amazon special.
Bezos studied engineering and computer science at Princeton, graduating magna cum laude. He then went to work doing IT for Wall Street . From beginning to end, he's been about expanding computer technology. He didn't build infrastructure in order to sell books, he used books and other easily shippable products to monetize a computer based distribution system. You may notice they sell a heck of a lot more than books - because books are an readily replaceable accessory to their actual business. That's why they don't write books, they buy them because books are not what makes Amazon special.
The idea for Amazon came to him while he was traveling a across the country and he heard that the supreme court ruled internet sellers don't have to collect sales tax. He decided to combine that with his skill at building large scale infrastructure and put together a mass market system selling stuff on the internet at a discount. What to sell using the system he designed? It should be valuable enough to ship. You don't sell concrete or soda online, shipping would be a problem. Electronics have high value per pound, but quickly lose value in the warehouse. The post office has a special extra low shipping rate for books, so books were good product to start with. The product was chosen to fit the distribution infrastructure. The infrastructure wasn't built to put his (nonexistent) bookstore online.
--checksum sounds like a winner. I may need to review man rsync, then see if we should be using that anywhere.
>OK, I haven't actually used btrfs (or any other fs with similar snapshot/CoW capabilities
Neither have I, but I don't think we're supposed to admit that on/. I think we're supposed to act as though we're experts on things we've never seen before, since this is Slashdot.
I have read the code for copy-on-write snapshots used by lvm and my understanding is that it's essentially similar.
I replied to the person who said today's deficit is half of six years ago. So I compared today to six years ago.
You wouldn't have a reading comprehension problem would you? No, not you.
Reagan won his election in a landslide because interest rates were around 21%, unemployment was 7.5%, and inflation was high. Reagan made some investments to cut interest rates in half, decrease taxes by $15,000 per family, slash inflation, reduce unemployment, and win the cold war.
Obama spent the same amount of money to - let his donors retire from their bogus solar companies. There's a difference between borrowing a mortgage of $100,000 to buy a house vs. throwing $100,000 in the camp fire. I have to believe you're smart enough to understand that.
I don't think it matters that their leadership _can_ do it. Do they _want_ to become an enterprise software company? If not, it's a distraction from the goal. If they want to be an electric car company, they should focus their energy on electric cars, not email, SAP, staplers sandwiches or anything else they _could_ build.
Disclaimer - Some have posted that SAP is CRAP and there is no reasonable alternative. I find it hard to believe that there is no off-the-shelf software to fit the need, but if so, that would justify building it, just because they have no reasonable alternative.
You have a point, but there's an important counter-point. The way the federal budget works for most projects, each year it's assumed that each project will get 104% of the spending it got last year. So in 1960 they approve $DUMB_IDEA. Fifty years later, $DUMB_IDEA is still eating away at your paycheck, with a budget ten times as high as the initial experiment. If you start with an across-the-board cut* and then look at which programs should have funding restored*, that means someone has to look at $DUMB_IDEA and either let it die or be responsible for continuing it.
Sure, some dumb ideas will stick around. Others were proposed by and for the Senate's most powerful member, Robert KKK Byrd, and kept going for decades because it was money for his state (or often, his ego). Now that Byrd is dead, some of his pork can finally be ended by the review process that will be required to expand the budget past the general cuts.
* "cut" in government parlance means an increase of 1% rather than 4% - still an increase.
"restored" meaning "set to grow quickly, rather than slowly"
"exactly meets the needs of the business" is important for some things, a huge waste of time and money for others. Those "some things" where it matters are generally the core competency of the business - what sets them apart from competitors. Google search needs a database that exactly meets their needs for searching a huge database. MySQL won't meet their need. For 99% of businesses, building a custom database engine would be stupid - MySQL, MS SQL, or Oracle would meet not only their current needs, but also their future needs.
Future needs can be a huge hidden expense for custom work; you've saddled the company with a requirement to build 2.0, 3,0, etc. on down the road if your business is built on something custom. So you should ask yourself "does the next company down the block have a similar need as we do?" If so, you and the company down the block should probably be sharing the development cost by both buying the same off-the-shelf software. You don't custom design your own trash cans, and most software is the same - yours should be about the same as the other guy uses.
If off-the-shelf software provides 95% of what you need but you need 5% custom, that's where FOSS is a perfect fit. You get 95% of it done, tested and ready to go, for free, and you just develop the 5% that you need special.
Given that they only spent a few months on it and don't have experience building broadly applicable SAP systems, we can be pretty certain you are correct in this statement:
> Their system is probably custom-tailored to their business processes. Not only would it not be appropriate for many other businesses...
It's probably still true if we change a few words:
> Their system is probably custom-tailored to their current business processes. Not only would it not be appropriate for many other businesses, including Tesla a year or two from now,...
Generally, you should build within your core competency, and buy generic systems for generic tasks. Tesla should design their own cars, especially electrical subsystems of the cars, but buy trashcans, spreadsheets, and SAP. Their SAP needs aren't that different from the next company down the road.
In a gray area, where you need something customized to your needs, but it's mostly the same as what other companies use, FOSS fits the bill. You get the 95% of common functionality free, then build the 5% that's unique to your needs.
With interrupts disabled latency in measured in nanoseconds. Of course for most applications, jitter matters more than latency. that is to say, if you know the latency is 100ns, you execute the instruction 100ns sooner. With USB, latency could be 100ns or 10,000ns, it's not consistent, so you can't control the timing.
Two reasons. One, on the desktop it's hard to get into the programming that gets kids excited. A "robot" that handles your pet feeding chore for you is way cooler than printing words on the screen.
When I was a kid, we still cooked pudding, which had to be stirred non-stop while it cooked . I built a machine that did the stirring.
Also, while Python is very useful, there are things you won't learn with Python. Like learning basic arithmetic before you use a calculator, anyone working with technology benefits from understanding what's going on at a lower level and the Pi encourages learning about bits and bytes and how they relate to real things happening in the real world - motors spinning, lasers flashing, etc.
Some people who have only worked at a high level don't think they'd gain much, but most who learn are glad they did. You CAN drive without knowing what brake pads are, but a professional driver who knows what happens when he presses the pedal will have a distinct advantage in the mountains, where brake pads can overheat and fail. If you know how your car works, you'll know if a noise means "pull over immediately" or "change the pads this month". Programming is the same. Understanding low level as well as high makes someone a much better Python /.Net / Perl programmer.
> Backups are rotated on two drives at least, and they shouldn't be both online at any time, in case of power surge or lightning, or even hacking of the backup server. > The only way I see to achieve this, with one backup site, is to have backup drive B physically disconnected while backup drive A is plugged.
> How could it be fully automated ?
If your backup facility isn't protected from lightning and you can only have one facility, that does make it harder. Given those requirements, I suppose one could use a relay on the power line. For the data line, I guess you'd need something that includes chips like http://www.mouser.com/Semiconductors/Switch-ICs/Analog-Switch-ICs/_/N-7590c/ and I good alternate path to ground.
You make some good points. I think this statement deserves some clarification because after you said it, you mentioned why it's not wholly true:
> This is not a problem with using btrfs snapshots to perform incremental backups -- it's a problem with incremental backups.
It's a problem with SOME approaches to incremental backups. As you also said:
> then do an rsync into that incremental backup with --ignore-times
That's one way to do incremental without the "corrupt forever" problem, it works as one type of validation. So incremental can be done without that problem.
> fakery... to make each incremental backup *look* like a full backup.
Indeed, btrfs snapshots and some similar approaches *look* more comprehensive than they are. It's a trap for the unwary.
Indeed btrfs checksums with a custom scrub job could detect some types of corruption, corruption that occurs after the backup. By default, the btrfs scrub will try to recover the bad block when it's read. For this use case, you need it to delete or at least touch() the file. That would be the "not without a solid validation regime anyway" part of my post.
iPhone's are for hipsters. OSX is certified UNIX running on rock solid, high performance hardware. Don't confuse the two.
I used Linux exclusively for fifteen years. I've contributed to many open source projects, including the Linux kernel, and I'm the maintainer of Linux::LVM and other projects. In other words, I'm a fan of Linux. From one fan of Linux to another, don't dismiss OSX just because the same company makes overpriced toys as well. It's a solid UNIX which will run all of your favorite FOSS software, and do it well.
You're saying locks are useless because they are used? I'm not sure what you're trying to say here. Locks and polygraphs can both be beaten. I say that doesn't make them useless.
> Iove how you are [were] a locksmith *and* a magician
As a kid, I got into magic. I studied the most famous magician of all, Houdini. Houdini was famous as an escape artist, he'd get out of handcuffs, locked boxes and jail cells. I studied to be like Houdini. That's how my weekend magician gigs lead directly to a short stint working as a locksmith. Fyi, almost all magicians are *and* a day job. A magic show is a 30 minute event on a Friday or Saturday. Magician's don't work 9-5 except the ONE guy with the CBS contract (Henning, Copperfield) and three in Vegas.
From locksmithing and "tricking people" (magic) I got into security, which is my long-term career.
Snapshots can certainly be part of a backup system, and the btrfs variety can be a convenient local "undo" step, similar to the undo in your editor. The way they function is not appropriate for offsite backups, though, not without a solid validation regime anyway.
The problem with btrfs snapshots is that if data is corrupted at any point, that data silently remains corrupted in future snapshots because they are copy-on-write. Suppose, for example, that your backup of your httpd.conf file gets corrupted at any point. If you take a snapshot, the "new" version is actually the old version - still corrupted. You then rsync from the source. Rsync looks at the mtime and because the file isn't marked as changed on the source, it doesn't get synced. All backups are still corrupted. For a litle while, you might have a copy from before the corruption, but eventually that ages out and you're left with nothing but corrupted data. I'm not theorizing on this point. I learned this lesson the hard way.
You've brought up some interesting topics. I bet you could ask some really good questions regarding the topics you mentioned. Instead, you chose to make ludicrous assumptions, assuming answers to the questions you could have asked.
Do you have any reason to make any of those assumptions? What makes you think some companies, such as a certain insurance company, don't run Clonebox on-site, mirroring between their facilities? Are you familiar with the datacenter choices available with Clonebox, including the underground nuclear bunker?
Do you think it's even POSSIBLE to offer that service at that price WITHOUT deduplication?
Do you somehow think you know anything at all about the media used?
> if it is possible to fool the polygraph it leaves no doubt that the polygraph is not scientific or useful
Your eyes can be fooled. Therefore they are not useful? Locks can be picked. Therefore they are not useful?
I used to work as a magician and a locksmith, so I can fool your eyes, and your locks. Now that you know your eyes can be fooled and are therefore useless, you're getting rid of them I guess?
If your eyes tell you that I just put your watch in my pocket, that's PROBABLY true. If a polygraph tells you that a stole your watch, it's probably right. Witnesses and polygraphs are about equally reliable.
"rsync at midnight". At 8:00 AM, discover that your filesystem got hosed at 10:00 PM, so you now have two copies of garbage.
Do not just sync periodically. Approximately everyone I've seen try that method got screwed in the end. They'd discover that they got rooted two weeks before, they'd overwritten an important file two days before, etc. You must ROTATE and then sync to be doing anything more than pretending that you have a backup. me.
The attributes of a good backup system:
Backups must be fully automatic, otherwise you'll stop doing them regularly.
Backups should be rotated. A midnight backup is useless if you are hacked at 11:55 PM, or discover a problem 2 days later. You must have access to older backups.
Backups must be offsite. Fires and burglars will take your backup if it is on site.
Backups must be accessible. As OP said, spending two weeks downloading your data isn't acceptable.
Backups must be tested. Our experience with web servers indicates that approximately 60% of backups provided by hosting providers don't actually work when you try to restore them
To meet all of the above requirements, we use an enterprise grade system called Clonebox. Other systems may be more applicable for home use.
I have three options I'll present for you. One matches your headline, one is cheap, and one is really, really solid.
The option that most matches your headline would be to use a WIFI NAS at the next door neighbor's house. Use any of the many good backup software packages. More on what a "good" backup system is in a moment.
Something I used to do was have two external drives. On Mondays, I'd switch out the drive in the house for the one in the car, which would go to work with me. The drawback to that is it's not fully automatic, so sometimes I'd forget or be in a rush. That leads us to the attributes of a good backup system:
Backups must be fully automatic, otherwise you'll stop doing them regularly.
Backups should be rotated. A midnight backup is useless if you are hacked at 11:55 PM, or discover a problem 2 days later. You must have access to older backups.
Backups must be offsite. Fires and burglars will take your backup if it is on site.
Backups must be accessible. As you said, spending two weeks downloading your data isn't acceptable.
Backups must be tested. Our experience with web servers indicates that approximately 60% of backups provided by hosting providers don't actually work when you try to restore them
To meet all of the above requirements, we use an enterprise grade system. It may be overkill for your needs, but then again the $8 / month version may be just what you want. It provides several offsite backups from different points in time and they are BOOTABLE. You can pull down a file or two, run a program or service remotely, or restore a full system.
3-4 Mbps to transfer 1TB is no good, as you said, but you actually have 200 Mbps available if you use the system we use. If you need the entire 1 TB, not just a small part of it, the whole 1TB bootable drive will be delivered to your front door within 12 hours. You may know the old saying "never underestimate the bandwidth of a station wagon full if tapes.". With a 1TB drive, the bandwidth of FedEX is over 200 Mbps.
What we use is called Clonebox. It's designed more for business, but it may either work for you, or give you some ideas.
> with wealthier students would try paying their teachers more(and so being able to get the best ones), while the schools that had poorer students
Instead of handing $10,000 of tax money to whichever school is nearest the student, the money goes to the school that the parents choose. It's thesame money being spent today. The only difference is that in order to get the money and students, schools need to not suck. We have a limited form of this where Ilive and it works well.
In fact, where I live not only do we have GOOD schools that students and parents can choose, they can choose the one that's best FOR THEIR CHILD. We have a high school for the arts, a collegiate high school where students graduate with an associate's degree, etc. All are tuition free - as public schools, they are financed by taxes
> raising the class size to 100
Schools here haven't tried that because if they did, students wouldn't choose that school. What they have done is ask students who are leaving "why are you choosing the other school?" They look at what works in the next school to see if it would work for them. Sometimes they decide to emulate the other school, sometimes it's different strokes for different folks. The science academy, for example, doesn't try to be just like the school for the arts.
I could post examples of questions created by instructional designers at the government agency I work for that are also horrible. Neither government nor private companies have a monopoly on stupid.
The big difference is that if a private company sells crap, purchasers can switch to a better vendor. If the government mandates garbage, you're screwed.
Slashdot Mirror
I guess you're unfamiliar with Bezos and unfamiliar with how and why Amazon began. Books are not what makes Amazon special. The idea of books came after Bezos designed the system that makes Amazon special.
Bezos studied engineering and computer science at Princeton, graduating magna cum laude.
He then went to work doing IT for Wall Street . From beginning to end, he's been about expanding computer technology. He didn't build infrastructure in order to sell books, he used books and other easily shippable products to monetize a computer based distribution system. You may notice they sell a heck of a lot more than books - because books are an readily replaceable accessory to their actual business. That's why they don't write books, they buy them because books are not what makes Amazon special.
The idea for Amazon came to him while he was traveling a across the country and he heard that the supreme court ruled internet sellers don't have to collect sales tax. He decided to combine that with his skill at building large scale infrastructure and put together a mass market system selling stuff on the internet at a discount. What to sell using the system he designed? It should be valuable enough to ship. You don't sell concrete or soda online, shipping would be a problem. Electronics have high value per pound, but quickly lose value in the warehouse. The post office has a special extra low shipping rate for books, so books were good product to start with. The product was chosen to fit the distribution infrastructure. The infrastructure wasn't built to put his (nonexistent) bookstore online.
--checksum sounds like a winner. I may need to review man rsync, then see if we should be using that anywhere.
>OK, I haven't actually used btrfs (or any other fs with similar snapshot/CoW capabilities
Neither have I, but I don't think we're supposed to admit that on /. I think we're supposed to act as though we're experts on things we've never seen before, since this is Slashdot.
I have read the code for copy-on-write snapshots used by lvm and my understanding is that it's essentially similar.
I replied to the person who said today's deficit is half of six years ago. So I compared today to six years ago.
You wouldn't have a reading comprehension problem would you? No, not you.
Reagan won his election in a landslide because interest rates were around 21%, unemployment was 7.5%, and inflation was high. Reagan made some investments to cut interest rates in half, decrease taxes by $15,000 per family, slash inflation, reduce unemployment, and win the cold war.
Obama spent the same amount of money to - let his donors retire from their bogus solar companies. There's a difference between borrowing a mortgage of $100,000 to buy a house vs. throwing $100,000 in the camp fire. I have to believe you're smart enough to understand that.
2008 deficit: $458 billion
2013 deficit: $680 billion
2014 deficit: $744 billion
> Maybe you should make it a personal mission to learn about this stuff instead of just repeat dumbass headlines?
Perhaps you should. Rather than venturing to comedycentral.com on that mission, may I suggest you start with http://www.treasury.gov/ and http://www.cbo.gov/
I don't think it matters that their leadership _can_ do it. Do they _want_ to become an enterprise software company?
If not, it's a distraction from the goal. If they want to be an electric car company, they should focus their energy on electric cars, not email, SAP, staplers sandwiches or anything else they _could_ build.
Disclaimer - Some have posted that SAP is CRAP and there is no reasonable alternative. I find it hard to believe that there is no off-the-shelf software to fit the need, but if so, that would justify building it, just because they have no reasonable alternative.
That's probably true, from the very little I know about SAP. Is there not a good off-the-shelf or customizable alternative?
You have a point, but there's an important counter-point. The way the federal budget works for most projects, each year it's assumed that each project will get 104% of the spending it got last year. So in 1960 they approve $DUMB_IDEA. Fifty years later, $DUMB_IDEA is still eating away at your paycheck, with a budget ten times as high as the initial experiment. If you start with an across-the-board cut* and then look at which programs should have funding restored*, that means someone has to look at $DUMB_IDEA and either let it die or be responsible for continuing it.
Sure, some dumb ideas will stick around. Others were proposed by and for the Senate's most powerful member, Robert KKK Byrd, and kept going for decades because it was money for his state (or often, his ego). Now that Byrd is dead, some of his pork can finally be ended by the review process that will be required to expand the budget past the general cuts.
* "cut" in government parlance means an increase of 1% rather than 4% - still an increase.
"restored" meaning "set to grow quickly, rather than slowly"
"exactly meets the needs of the business" is important for some things, a huge waste of time and money for others. Those "some things" where it matters are generally the core competency of the business - what sets them apart from competitors. Google search needs a database that exactly meets their needs for searching a huge database. MySQL won't meet their need. For 99% of businesses, building a custom database engine would be stupid - MySQL, MS SQL, or Oracle would meet not only their current needs, but also their future needs.
Future needs can be a huge hidden expense for custom work; you've saddled the company with a requirement to build 2.0, 3,0, etc. on down the road if your business is built on something custom. So you should ask yourself "does the next company down the block have a similar need as we do?" If so, you and the company down the block should probably be sharing the development cost by both buying the same off-the-shelf software. You don't custom design your own trash cans, and most software is the same - yours should be about the same as the other guy uses.
If off-the-shelf software provides 95% of what you need but you need 5% custom, that's where FOSS is a perfect fit.
You get 95% of it done, tested and ready to go, for free, and you just develop the 5% that you need special.
Given that they only spent a few months on it and don't have experience building broadly applicable SAP systems, we can be pretty certain you are correct in this statement:
> Their system is probably custom-tailored to their business processes. Not only would it not be appropriate for many other businesses ...
It's probably still true if we change a few words:
> Their system is probably custom-tailored to their current business processes. Not only would it not be appropriate for many other businesses, including Tesla a year or two from now, ...
Generally, you should build within your core competency, and buy generic systems for generic tasks.
Tesla should design their own cars, especially electrical subsystems of the cars, but buy trashcans, spreadsheets, and SAP.
Their SAP needs aren't that different from the next company down the road.
In a gray area, where you need something customized to your needs, but it's mostly the same as what other companies use, FOSS fits the bill. You get the 95% of common functionality free, then build the 5% that's unique to your needs.
I suppose this is the suckcessor to Friday's submission.
With interrupts disabled latency in measured in nanoseconds. Of course for most applications, jitter matters more than latency. that is to say, if you know the latency is 100ns, you execute the instruction 100ns sooner. With USB, latency could be 100ns or 10,000ns, it's not consistent, so you can't control the timing.
Two reasons. One, on the desktop it's hard to get into the programming that gets kids excited. A "robot" that handles your pet feeding chore for you is way cooler than printing words on the screen.
When I was a kid, we still cooked pudding, which had to be stirred non-stop while it cooked . I built a machine that did the stirring.
Also, while Python is very useful, there are things you won't learn with Python. Like learning basic arithmetic before you use a calculator, anyone working with technology benefits from understanding what's going on at a lower level and the Pi encourages learning about bits and bytes and how they relate to real things happening in the real world - motors spinning, lasers flashing, etc.
Some people who have only worked at a high level don't think they'd gain much, but most who learn are glad they did. You CAN drive without knowing what brake pads are, but a professional driver who knows what happens when he presses the pedal will have a distinct advantage in the mountains, where brake pads can overheat and fail. If you know how your car works, you'll know if a noise means "pull over immediately" or "change the pads this month". Programming is the same. Understanding low level as well as high makes someone a much better Python / .Net / Perl programmer.
> Backups are rotated on two drives at least, and they shouldn't be both online at any time, in case of power surge or lightning, or even hacking of the backup server.
> The only way I see to achieve this, with one backup site, is to have backup drive B physically disconnected while backup drive A is plugged.
> How could it be fully automated ?
If your backup facility isn't protected from lightning and you can only have one facility, that does make it harder.
Given those requirements, I suppose one could use a relay on the power line. For the data line, I guess you'd need something
that includes chips like http://www.mouser.com/Semiconductors/Switch-ICs/Analog-Switch-ICs/_/N-7590c/ and I good alternate path to ground.
You make some good points. I think this statement deserves some clarification because after you said it, you mentioned why it's not wholly true:
> This is not a problem with using btrfs snapshots to perform incremental backups -- it's a problem with incremental backups.
It's a problem with SOME approaches to incremental backups. As you also said:
> then do an rsync into that incremental backup with --ignore-times
That's one way to do incremental without the "corrupt forever" problem, it works as one type of validation.
So incremental can be done without that problem.
> fakery ... to make each incremental backup *look* like a full backup.
Indeed, btrfs snapshots and some similar approaches *look* more comprehensive than they are. It's a trap for the unwary.
Indeed btrfs checksums with a custom scrub job could detect some types of corruption, corruption that occurs after the backup.
By default, the btrfs scrub will try to recover the bad block when it's read. For this use case, you need it to delete or at least touch() the file.
That would be the "not without a solid validation regime anyway" part of my post.
iPhone's are for hipsters. OSX is certified UNIX running on rock solid, high performance hardware. Don't confuse the two.
I used Linux exclusively for fifteen years. I've contributed to many open source projects, including the Linux kernel, and I'm the maintainer of Linux::LVM and other projects. In other words, I'm a fan of Linux. From one fan of Linux to another, don't dismiss OSX just because the same company makes overpriced toys as well. It's a solid UNIX which will run all of your favorite FOSS software, and do it well.
You're saying locks are useless because they are used? I'm not sure what you're trying to say here. Locks and polygraphs can both be beaten. I say that doesn't make them useless.
> Iove how you are [were] a locksmith *and* a magician
As a kid, I got into magic. I studied the most famous magician of all, Houdini. Houdini was famous as an escape artist, he'd get out of handcuffs, locked boxes and jail cells. I studied to be like Houdini. That's how my weekend magician gigs lead directly to a short stint working as a locksmith. Fyi, almost all magicians are *and* a day job. A magic show is a 30 minute event on a Friday or Saturday. Magician's don't work 9-5 except the ONE guy with the CBS contract (Henning, Copperfield) and three in Vegas.
From locksmithing and "tricking people" (magic) I got into security, which is my long-term career.
Snapshots can certainly be part of a backup system, and the btrfs variety can be a convenient local "undo" step, similar to the undo in your editor. The way they function is not appropriate for offsite backups, though, not without a solid validation regime anyway.
The problem with btrfs snapshots is that if data is corrupted at any point, that data silently remains corrupted in future snapshots because they are copy-on-write. Suppose, for example, that your backup of your httpd.conf file gets corrupted at any point. If you take a snapshot, the "new" version is actually the old version - still corrupted. You then rsync from the source. Rsync looks at the mtime and because the file isn't marked as changed on the source, it doesn't get synced. All backups are still corrupted. For a litle while, you might have a copy from before the corruption, but eventually that ages out and you're left with nothing but corrupted data. I'm not theorizing on this point. I learned this lesson the hard way.
You've brought up some interesting topics. I bet you could ask some really good questions regarding the topics you mentioned. Instead, you chose to make ludicrous assumptions, assuming answers to the questions you could have asked.
Do you have any reason to make any of those assumptions? What makes you think some companies, such as a certain insurance company, don't run Clonebox on-site, mirroring between their facilities? Are you familiar with the datacenter choices available with Clonebox, including the underground nuclear bunker?
Do you think it's even POSSIBLE to offer that service at that price WITHOUT deduplication?
Do you somehow think you know anything at all about the media used?
That's a higher quality response than often found here on /. when someone refutes a post, thank you.
> if it is possible to fool the polygraph it leaves no doubt that the polygraph is not scientific or useful
Your eyes can be fooled. Therefore they are not useful? Locks can be picked. Therefore they are not useful?
I used to work as a magician and a locksmith, so I can fool your eyes, and your locks. Now that you know your eyes can be fooled and are therefore useless, you're getting rid of them I guess?
If your eyes tell you that I just put your watch in my pocket, that's PROBABLY true. If a polygraph tells you that a stole your watch, it's probably right. Witnesses and polygraphs are about equally reliable.
"rsync at midnight". At 8:00 AM, discover that your filesystem got hosed at 10:00 PM, so you now have two copies of garbage.
Do not just sync periodically. Approximately everyone I've seen try that method got screwed in the end. They'd discover that they got rooted two weeks before, they'd overwritten an important file two days before, etc. You must ROTATE and then sync to be doing anything more than pretending that you have a backup.
me.
The attributes of a good backup system:
Backups must be fully automatic, otherwise you'll stop doing them regularly.
Backups should be rotated. A midnight backup is useless if you are hacked at 11:55 PM, or discover a problem 2 days later. You must have access to older backups.
Backups must be offsite. Fires and burglars will take your backup if it is on site.
Backups must be accessible. As OP said, spending two weeks downloading your data isn't acceptable.
Backups must be tested. Our experience with web servers indicates that approximately 60% of backups provided by hosting providers don't actually work when you try to restore them
To meet all of the above requirements, we use an enterprise grade system called Clonebox. Other systems may be more applicable for home use.
I have three options I'll present for you. One matches your headline, one is cheap, and one is really, really solid.
The option that most matches your headline would be to use a WIFI NAS at the next door neighbor's house. Use any of the many good backup software packages. More on what a "good" backup system is in a moment.
Something I used to do was have two external drives. On Mondays, I'd switch out the drive in the house for the one in the car, which would go to work with me. The drawback to that is it's not fully automatic, so sometimes I'd forget or be in a rush. That leads us to the attributes of a good backup system:
Backups must be fully automatic, otherwise you'll stop doing them regularly.
Backups should be rotated. A midnight backup is useless if you are hacked at 11:55 PM, or discover a problem 2 days later. You must have access to older backups.
Backups must be offsite. Fires and burglars will take your backup if it is on site.
Backups must be accessible. As you said, spending two weeks downloading your data isn't acceptable.
Backups must be tested. Our experience with web servers indicates that approximately 60% of backups provided by hosting providers don't actually work when you try to restore them
To meet all of the above requirements, we use an enterprise grade system. It may be overkill for your needs, but then again the $8 / month version may be just what you want. It provides several offsite backups from different points in time and they are BOOTABLE. You can pull down a file or two, run a program or service remotely, or restore a full system.
3-4 Mbps to transfer 1TB is no good, as you said, but you actually have 200 Mbps available if you use the system we use. If you need the entire 1 TB, not just a small part of it, the whole 1TB bootable drive will be delivered to your front door within 12 hours. You may know the old saying "never underestimate the bandwidth of a station wagon full if tapes.". With a 1TB drive, the bandwidth of FedEX is over 200 Mbps.
What we use is called Clonebox. It's designed more for business, but it may either work for you, or give you some ideas.
Those are both problems. SOLVED problems.
> with wealthier students would try paying their teachers more(and so being able to get the best ones), while the schools that had poorer students
Instead of handing $10,000 of tax money to whichever school is nearest the student, the money goes to the school that the parents choose. It's thesame money being spent today. The only difference is that in order to get the money and students, schools need to not suck. We have a limited form of this where Ilive and it works well.
In fact, where I live not only do we have GOOD schools that students and parents can choose, they can choose the one that's best FOR THEIR CHILD. We have a high school for the arts, a collegiate high school where students graduate with an associate's degree, etc. All are tuition free - as public schools, they are financed by taxes
> raising the class size to 100
Schools here haven't tried that because if they did, students wouldn't choose that school. What they have done is ask students who are leaving "why are you choosing the other school?" They look at what works in the next school to see if it would work for them. Sometimes they decide to emulate the other school, sometimes it's different strokes for different folks. The science academy, for example, doesn't try to be just like the school for the arts.
>
I could post examples of questions created by instructional designers at the government agency I work for that are also horrible. Neither government nor private companies have a monopoly on stupid.
The big difference is that if a private company sells crap, purchasers can switch to a better vendor. If the government mandates garbage, you're screwed.