Now, if you recall what happened with AWS in April, they had a low-bandwidth management network that all of a sudden had all primary EBS API traffic shunted to it. This was caused by a human flipping a network switch when they shouldn't have. Something like this is not something that happens all the time, has little, if any diagnosable features, is not well-defined to have a proper workflow attached to it, and needs human engineers to correct. This is an example of a complex, large-scale problem.
I wonder when this army of automated-problem-fixing engines will encounter a corner case its masters never considered and how it will react.
I give the ops guys at Facebook a lot of credit for managing such a gigantic workload with just a (relatively) few, very smart, people. Amazon also has a lot of smart people who have been working on EBS (in one form or another) since before Facebook was founded. These systems just interact in unpredictable ways when they get out of their comfort zone.
Systems so complicated they require self-managing management systems are going to have some interesting failure modes, to say the least.
Guess who was involved in said application. TEPCO.
I certainly hope it gets denied. Not a single person from that company should ever be allowed to touch a reactor again.
Actually, it is uncommon for someone to modify their automobile.
Only 3.5% of people will use non-OEM parts if they are given an equal-cost choice, so the number who would go for an out-of-spec non-OEM replacement/upgrade is even smaller.
Basically, you are trying to say that something that occurs with less frequency than coming across someone running Linux on their Desktop is actually common. Bzzzzt.
But would the same 8 spindles handle the 20,000 IO/sec rate that this single drive manages?
Not everyone is bound by the raw sequential transfer rate. In fact, almost no one is.
These numbers don't really represent that much. They are for U.S. Retail sales. Since Apple is very dominant in the (tiny) retail computer sales industry, its not a shocker that they have high market share in a slice of that market.
If you were to count BTO computers sold over phone or internet in the U.S. Apple's market share would drop. Add the rest of the world and Apple's market share shrinks even more.
That said, Apple is gaining speed and is only going to be selling more computers for the foreseeable future.
One nice thing about electric cars is that they will typically be charging at night.
Power demands are much lower at night, so a population charging electric cars at night might allow us to make more efficient use of the grid all day long, instead of building it to handle a peak load it only sees 2 hours a day.
If you'd look you would see that health and human services eat up over 50% of the entire budget. We spend nearly half the defense budget on servicing our national debt.
To think that 30% couldn't be cut from that budget is insane.
99.9% of the time a user is not bottlenecked on throughput or overall bandwidth. They are hamstrung by Latency. This is where SSDs really shine. They can retrieve any piece of data from anywhere in their storage in 1/100 or 1/1000th the time it would take a standard hard drive to do the same.
It is this random access low latency performance that really improves the experience for the users.
Wear leveling is handled in the firmware on the flash storage device itself.
It maintains a table which matches the logical block addresses to physical cell locations on the flash memory. The mapping between the logical and physical locations in handled by the wear leveling algorithms, which you can not inspect or influence from the devices data interface. Thus no program running on the host device can create a pathological case of wearing out a specific cell. Only bad firmware could do that.
Certainly they are using wear leveling. They are probably (information is thin) also over allocating storage, such that 10-15% can fail before impacting the advertised free space on the device.
In reality you will see your full 640GB of storage, which you could write 61PB of data to the very first sector on the disk over and over and never experience any issues. Before the last 'extra' block is used up, you'll get an alert and replace the device.
Perhaps they did not have any clue as to the 'why' but thought that the data was interesting enough to warrant further research.
This is a great way to get working on the 'why', as without this paper no one would be looking at it.
This is one way that science is done. They probably postulated that the alignment of galaxies would be random, and when they tested this hypothesis they found that the data did not match. Publishing that result so that others can start working on it is the next step in this process.
(Note: not every oil company is diversifying into renewables. Some dinosaurs, like Exxon-Mobil, resist it like the plague. But many are.)
Well, if Exxon keeps piling up the profits (and they are) in the short term, then 10 years down the road they can buy whichever company develops interesting renewable energy.
By concentrating on exploiting oil at this moment, they can use capital down the road to help them migrate to a longer-term renewable energy source. Smart really.
Maybe they will just start collecting stats on your power usage. If they did so, they could tell if you were a likely consumer of electronic goods, discern the times of day when you are most active, have a reason to pull full credit information on every subscriber to their power utility, etc. Google could apply brainpower to something basic like electrical supply and leverage that to add to their data models on you, thus making ads served in other venues (web, tv, radio) more closely matched.
Just one train of thought on why Google may possibly make more money off running an electricity company that a standard Utility would.
nothing new said here, just dont assume that because of US's huge and not so densely populated area train travel would work. it does.
If it would work (economically) then why doesn't it happen? As part of a free market society, companies should be rushing to provide this service if it can be done profitably.
The fact that it is not being done shows that either 1) it just recently became possible to do it profitably and people are working on it right now, or 2) its not profitable.
My money is on 2.
I hope that the US can bypass the whole rail/maglev idea and jump right to the flying car stage. Maybe sometime before I die at least.
I said sadly in reference to the 'dream' of rail transport being squashed in the US. I am a big fan of the American lifestyle, don't get me wrong.It just does not lend itself to rail transport (or mass transit in general).
Yeah, if all out population in the US was situated with very high density in an almost straight line, rail would be an option.
Sadly, the American Dream includes owning a Home, with a yard and all that fun stuff. This means that we don't have the population densities outside of a few major metropolitan areas to support rail travel.
The other downside is that our population centers are _far_ away from each other. People from Asian or European countries just don't understand how much space lies between American cities.
The United States today does not have the economics going for rail transport that some other countries have. That is why we don't have the rail transport systems that other countries have. It doesn't make economical sense.
As long as distribution of goods is tied to the money one can obtain for it, then artificially creating shortage through the systematic destruction of that good is an excellent idea.
Only until you succeed in raising the price of _your_ products past the point at which others can supply a sufficient volume of goods at a high enough price to turn a profit.
Unless you have a complete and total monopoly (utility companies for example), there is always a chance that others will see the high prices you are charging, and move into your area. The glory of capitalism is that it provides lots of opportunity for companies to move into new product spaces quickly and easily to balance out those who would profit via destruction.
Slashdot Mirror
Now, if you recall what happened with AWS in April, they had a low-bandwidth management network that all of a sudden had all primary EBS API traffic shunted to it. This was caused by a human flipping a network switch when they shouldn't have. Something like this is not something that happens all the time, has little, if any diagnosable features, is not well-defined to have a proper workflow attached to it, and needs human engineers to correct. This is an example of a complex, large-scale problem.
I wonder when this army of automated-problem-fixing engines will encounter a corner case its masters never considered and how it will react.
I give the ops guys at Facebook a lot of credit for managing such a gigantic workload with just a (relatively) few, very smart, people. Amazon also has a lot of smart people who have been working on EBS (in one form or another) since before Facebook was founded. These systems just interact in unpredictable ways when they get out of their comfort zone.
Systems so complicated they require self-managing management systems are going to have some interesting failure modes, to say the least.
Guess who was involved in said application. TEPCO. I certainly hope it gets denied. Not a single person from that company should ever be allowed to touch a reactor again.
Actually, it is uncommon for someone to modify their automobile. Only 3.5% of people will use non-OEM parts if they are given an equal-cost choice, so the number who would go for an out-of-spec non-OEM replacement/upgrade is even smaller. Basically, you are trying to say that something that occurs with less frequency than coming across someone running Linux on their Desktop is actually common. Bzzzzt.
Actually the T2 currently scales to 4 sockets (the T5440) and has nothing standing in the way of scaling farther than that.
But would the same 8 spindles handle the 20,000 IO/sec rate that this single drive manages? Not everyone is bound by the raw sequential transfer rate. In fact, almost no one is.
If you were to count BTO computers sold over phone or internet in the U.S. Apple's market share would drop. Add the rest of the world and Apple's market share shrinks even more.
That said, Apple is gaining speed and is only going to be selling more computers for the foreseeable future.
Power demands are much lower at night, so a population charging electric cars at night might allow us to make more efficient use of the grid all day long, instead of building it to handle a peak load it only sees 2 hours a day.
pi*r^2
2007 Federal Budget
If you'd look you would see that health and human services eat up over 50% of the entire budget. We spend nearly half the defense budget on servicing our national debt.
To think that 30% couldn't be cut from that budget is insane.
No One Cares.
It is this random access low latency performance that really improves the experience for the users.
It maintains a table which matches the logical block addresses to physical cell locations on the flash memory. The mapping between the logical and physical locations in handled by the wear leveling algorithms, which you can not inspect or influence from the devices data interface. Thus no program running on the host device can create a pathological case of wearing out a specific cell. Only bad firmware could do that.
Certainly they are using wear leveling. They are probably (information is thin) also over allocating storage, such that 10-15% can fail before impacting the advertised free space on the device. In reality you will see your full 640GB of storage, which you could write 61PB of data to the very first sector on the disk over and over and never experience any issues. Before the last 'extra' block is used up, you'll get an alert and replace the device.
Cosidering that this drive is 640GB, that means you would need to write somehwere in the region of 61 PETABYTES of information.
You'd have to write to the drive at a perfect 800 MB/s for 941 days to hit that mark.
It could last as long as 30 years, at full write speed of 800 MB/s if it can handle 1M writes per cell.
At the end of the day, semiconductors this large and high quality are certainly better than tiny bits of rust on rapidly spinning platters.
Commercial products in the high-end flash space are promising 500,000+ writes.
We are not talking about glorified thumb-drive flash memory here, but decent chips with good wear leveling and high quality construction.
This is a great way to get working on the 'why', as without this paper no one would be looking at it.
This is one way that science is done. They probably postulated that the alignment of galaxies would be random, and when they tested this hypothesis they found that the data did not match. Publishing that result so that others can start working on it is the next step in this process.
I am sure that they have some gigantic UNIX machines rolling around in datacenters doing billing and such.
Use the best tool for the job. In some places, even at Apple, that means using Windows when its prudent.
Well, if Exxon keeps piling up the profits (and they are) in the short term, then 10 years down the road they can buy whichever company develops interesting renewable energy.
By concentrating on exploiting oil at this moment, they can use capital down the road to help them migrate to a longer-term renewable energy source. Smart really.
Just one train of thought on why Google may possibly make more money off running an electricity company that a standard Utility would.
North Koreans, Chinese and other assorted countries will just tell the US to shove it when they come looking for these people.
Since the technology is deployed worldwide, it can be hacked world wide. No amount of extra laws in the US is going to stop the next DVD Jon.
I still have an Mac IIfx. It even boots still.
And at the time, it was truly a wicked fast machine.
If it would work (economically) then why doesn't it happen? As part of a free market society, companies should be rushing to provide this service if it can be done profitably.
The fact that it is not being done shows that either 1) it just recently became possible to do it profitably and people are working on it right now, or 2) its not profitable.
My money is on 2.
I hope that the US can bypass the whole rail/maglev idea and jump right to the flying car stage. Maybe sometime before I die at least.
I said sadly in reference to the 'dream' of rail transport being squashed in the US. I am a big fan of the American lifestyle, don't get me wrong.It just does not lend itself to rail transport (or mass transit in general).
Sadly, the American Dream includes owning a Home, with a yard and all that fun stuff. This means that we don't have the population densities outside of a few major metropolitan areas to support rail travel.
The other downside is that our population centers are _far_ away from each other. People from Asian or European countries just don't understand how much space lies between American cities.
The United States today does not have the economics going for rail transport that some other countries have. That is why we don't have the rail transport systems that other countries have. It doesn't make economical sense.
Only until you succeed in raising the price of _your_ products past the point at which others can supply a sufficient volume of goods at a high enough price to turn a profit.
Unless you have a complete and total monopoly (utility companies for example), there is always a chance that others will see the high prices you are charging, and move into your area. The glory of capitalism is that it provides lots of opportunity for companies to move into new product spaces quickly and easily to balance out those who would profit via destruction.