Seldom (if ever) does a single job get all the processors. I think I only saw it when we were running LINPACK to see if we were hitting our numbers. That job runs for quite some time, and even when running on all 1000+ processors on our BlueGene I never once saw it fail due to hardware issues. In fact, in over a year I think I only saw one instance of a job having to be re-run due to hardware failures. I can't say my knowledge is exhaustive, but my personal experience is that the BlueGene supercomputer series has an excellent hardware availability record.
Go ahead - unify all the linuces under one big top! One Linux to rule them all, and all that.
Of course, the millisecond that happens, the thing'll fracture into a dozen or more different, uh, "tiers" (think: Basic, Home, Premium, Ultimate, or some junk like that). Next (but still well before 1000ms has passed), every Linux developer worth his salt will immediately create his or her own, uh, "customizations" - because ONE SIZE DOESN'T FIT ALL.
So - approximately one second after unifying all Linuces as one Linux, it'll immediately breed by fission - not into two but literally dozens of new, uh, . . . (can't call 'em distros any more. Custom Linux versions?).
Higher than you're guessing. I've worked on BlueGene/L, BlueGene/S and was involved in some of the development on BlueGene/P. All of these systems have an incredibly agressive monitoring mechanism - voltages, temperatures, fan speeds, as well as half a dozen other hardware categories are monitored at the component level and the data stored in a database where it is analyzed to ensure that the system as a whole IS operational and stays that way.
But thank you for pointing out that the architecture is inherently fault-tolerant. When submitting a job to massively parallel machines like this, one of the options presented is how many cores/how much memory for this job (as well as many other performance affecting options such as internal network topology). A bad core can be "skipped" in much the same way as a bad sector on a hard drive; except that in this case, it's possible to repair/replace the bad sector on the fly.
Let me know when widespread adoption seems likely.
Let me know when widespread support is available.
This is one of those cases where theory and practice differ. In theory, I'd love to wait until some absolutely uncrackable/fast/compact/available technology makes securing DNS possible. In the interim, this isn't the time to go back to square one and start over.
Of course, since DNScurve will never need a successor, of course it'll be worth the wait. Obviously, DNSSEC will have a successor and so we should just not bother and stick with good ol' DNS until DNScurve has wide enough adoption to make migrating work.
Uh, given that DNSSEC has taken nearly a decade to get here, how long will it be for DNScurve?
In theory, there's no difference between theory and practice. In practice . . .
That seems to be the crux of his arguement against DNSSEC - that RSA is broken (or soon to be broken).
Okay, let me restate the problem - should we implement a mechanism which is already available and well understood, as well as generally accepted as secure (Mr. Berstein's assertion that RSA1024 is broken to the contrary), or should we implement it with a technology which can't be nearly as mature (not so much for its newness, but rather for its lack of broad acceptance/use)?
You're right - let's pick the shiniest technology on the shelf, we all know that elliptic curve encryption is faster, smaller and uncrackable, right? Uh, it is uncrackable, right? 'Cuz, you see, RSA1024 is uncrackable, or so I was once told. Now Dr. Berstein says it ain't, but elliptical curve encryption is. Funny thing - if RSA1024 is more than enough to secure my bank transactions, why wouldn't I trust it with my DNS queries?
At just a moment when the internet at large needs to standardize on secure mechanisms, he has to gratuitously add another potential standard to the mix, increasing the difficulty of getting anything done.
If RSA were not considered computationally secure, I might applaud his intent to provide "a better mousetrap". However, since RSA is (to the best of my knowledge) still considered secure, his elliptical cryptography based version of DNS brings NOTHING of value to the table, and only serves to complicate what should otherwise be a reasonably straightforward proposition - that of migrating the internet's DNS servers to a secure DNS implementation.
So it can't be pornography to me until I see it (Schroedinger's kitty porn? Or was that a snuff film?).
I'm over eighteen - let me collapse the wave function for myself, willya? Oh, and someone please let the damned cat out of the box - the meowing is driving me nuts!
Y'know, I understand the "Great (Fire)Wall of China". Shouldn't this be the "Australian Great Barrier Firewall"? And - isn't that in danger of being destroyed by people poking and prodding at it, punching holes in it, etc.?
MicroSoft has been pretty clear that Windows 7 is not a major shift away from Vista. As an example: Win 3.1 -> Win95 was a heart transplant, Win95 -> Win98 was a facelift.
Windows Vista -> Windows 7 is another facelift. MicroSoft has been mostly honest about this. Yes, there have been conflicting rumors, some of which originated with MicroSoft personnel, but the source doesn't necessarily guarantee the accuracy of the information. MicroSoft's official line on Windows 7 is to quietly admit that it isn't a major code revision - in fact, I heard one rumor that they were planning to release it under the name "Windows 6.3" instead of the codename "Windows 7", to explicitly make it (sort of) clear that this is a minor revision, not a major one.
Not that I think Windows Vista is worth the DVD's it ships on, but the folks here might consider giving MicroSoft a break. My opinion of MicroSoft remains pretty low, but in this I think they're actually trying to do something right. They're not claiming to have invented a new paradigm or anything stupid like that, they're just saying that there's a new version of the Windows operating system coming soon.
That's right - let the Governments of the world fix the internet by legislation; after all, we all know how well the government understands the tubes of the intarweb. Perhaps Al Gore could be tapped to spearhead this incredibly important piece of legislation.
to incredibly high magnetic fields combined with high rotational velocities . ..
So . . . has any pulsar ever been gobbled up by a microscopic black hole, or turned into strange matter (stranglets), or caused a quantum vacuum explosion? No?
Now, if they'd said they were going to incorporate a Peltier device on the chip die to let 'em run heavily overclocked and ice-cold, I might've fallen for it . . .
In court, merely argue that other websites are also hyperlinking to, ah, the plaintiff's website and that this is not a matter of trademark/copyright protection, but rather a transparent attempt at a discriminatory practice. Countersue on that basis.
Legal nonsense, but when coupled with an honest, good faith campaign to ensure that no one creates links to their website, well . . . when even the Goog can't display your links, what's the point of having a web presence?
The only valid reason I can see for Google to leave the tag "beta" stuck on everything is so that when competitors (think: M$) give 'em a dig, such as "Oh, but Google Mail can't do {blah}", the PHB's at Google can immediately answer back "Well, Google Mail is still in beta. We're looking at implementation of {blah}, and expect it to be up and running shortly."
I can also see this in court - "Your honor, we didn't reverse engineer functionality {blah} - our system is still in beta. We were planning to implement {blah} all along. We didn't steal the idea from so-and-so."
Do no evil? Hmmm . . . how 'bout "Free the beast". Has anybody here looked at the price of Google stock lately? I use their tools regularly (frankly, I'm glad as hell that Google exists), but I don't fool myself. They're definitely in it for the money. Then again, so am I.
Hellfire, the government could issue an RSA code to every citizen and publish the public keys in a phone book. The government could even provide the necessary software to make it work. It'd be secure - that's the beauty of public key encryption systems such as RSA or knapsack. But it'll never happen. Nobody wants it. Nobody wants to pay for it.
This legislation will force industry to develop and pay for it, regardless of whether the customers want it or not. Yes, we all want encryption on everything; but an overwhelming majority of computer users don't care enough to actually do anything, even though it would only take a bit of time and effort. Now, what happens when your bank send you your private encryption key and instructions? Most recipients will either delete or (at best) ignore the key. Later that month imagine their anger when their bank statement is encrypted and they have no idea how to decrypt it? Or do you really get the impression that the average American (Nevadan?) consumer is intelligent enough to implement, say, GPG? If so, do you think the average consumer is energetic enough to do so?
Leave this job up to market forces - the free-enterprise economy is infinitely more responsive to the needs and wants of the average consumer than is the Federal or even any of the State governments.
While I'll grant you that businesses should absolutely use encryption or some other mechanism to protect sensative information, legislation isn't the solution.
Consider - if a bank sent new ATM cards with the pin in the same envelope as the card, most consumers would go immediately berserk. The institution in question would rapidly see an erosion of their customer base, as well as being found liable for any losses incurred by people who had their mail intercepted by thieves.
That same bank can blithely send out e-mails with user account names, numbers and passwords all in one convenient, easy-to-sniff package and nobody gets upset. How often has anybody here clicked on "forgot my username/password" only to get a nice, convenient clickable link which allows unfettered access to private, smooth, creamy soft personal information? The solution isn't for the government to legislate the use of encryption; rather, it's a matter for market pressure. Let enough people become unhappy over the cavelier treatment their personal information garners from a corporation - they'll vote with their wallets, if they are once educated regarding the situation.
That last phrase is the hard part though, isn't it?
How do they keep 'em patched, I wonder? Must be some kind of clustering scheme, since we all know that most Windows patches require rebooting, yes?
Yeah, give me enough hardware and I can make almost any OS (MS-DOS anyone) do whatever is needed in a multitasking server environment. Then again, Microsoft is quite famous for delivering software which requires (b)leeding edge hardware to run adequately.
In plain sight of the public, which might just barely conceivably still have sufficient intelligence and strength of will to stop you, or quietly, unobtrusively, all-but-unnoticed in the shadows?
"The price of Freedom is eternal vigilance." - who said that again?
Slashdot Mirror
Seldom (if ever) does a single job get all the processors. I think I only saw it when we were running LINPACK to see if we were hitting our numbers. That job runs for quite some time, and even when running on all 1000+ processors on our BlueGene I never once saw it fail due to hardware issues. In fact, in over a year I think I only saw one instance of a job having to be re-run due to hardware failures. I can't say my knowledge is exhaustive, but my personal experience is that the BlueGene supercomputer series has an excellent hardware availability record.
Of course, the millisecond that happens, the thing'll fracture into a dozen or more different, uh, "tiers" (think: Basic, Home, Premium, Ultimate, or some junk like that). Next (but still well before 1000ms has passed), every Linux developer worth his salt will immediately create his or her own, uh, "customizations" - because ONE SIZE DOESN'T FIT ALL.
So - approximately one second after unifying all Linuces as one Linux, it'll immediately breed by fission - not into two but literally dozens of new, uh, . . . (can't call 'em distros any more. Custom Linux versions?).
But thank you for pointing out that the architecture is inherently fault-tolerant. When submitting a job to massively parallel machines like this, one of the options presented is how many cores/how much memory for this job (as well as many other performance affecting options such as internal network topology). A bad core can be "skipped" in much the same way as a bad sector on a hard drive; except that in this case, it's possible to repair/replace the bad sector on the fly.
Let me know when widespread support is available.
This is one of those cases where theory and practice differ. In theory, I'd love to wait until some absolutely uncrackable/fast/compact/available technology makes securing DNS possible. In the interim, this isn't the time to go back to square one and start over.
Of course, since DNScurve will never need a successor, of course it'll be worth the wait. Obviously, DNSSEC will have a successor and so we should just not bother and stick with good ol' DNS until DNScurve has wide enough adoption to make migrating work.
Uh, given that DNSSEC has taken nearly a decade to get here, how long will it be for DNScurve?
In theory, there's no difference between theory and practice. In practice . . .
Okay, let me restate the problem - should we implement a mechanism which is already available and well understood, as well as generally accepted as secure (Mr. Berstein's assertion that RSA1024 is broken to the contrary), or should we implement it with a technology which can't be nearly as mature (not so much for its newness, but rather for its lack of broad acceptance/use)?
You're right - let's pick the shiniest technology on the shelf, we all know that elliptic curve encryption is faster, smaller and uncrackable, right? Uh, it is uncrackable, right? 'Cuz, you see, RSA1024 is uncrackable, or so I was once told. Now Dr. Berstein says it ain't, but elliptical curve encryption is. Funny thing - if RSA1024 is more than enough to secure my bank transactions, why wouldn't I trust it with my DNS queries?
A personal opinion, that. YMMV.
If RSA were not considered computationally secure, I might applaud his intent to provide "a better mousetrap". However, since RSA is (to the best of my knowledge) still considered secure, his elliptical cryptography based version of DNS brings NOTHING of value to the table, and only serves to complicate what should otherwise be a reasonably straightforward proposition - that of migrating the internet's DNS servers to a secure DNS implementation.
I'm over eighteen - let me collapse the wave function for myself, willya? Oh, and someone please let the damned cat out of the box - the meowing is driving me nuts!
What did Intel do to make it suitable?
Otherwise, how will they ever get the Farscape module into orbit?
I'd like to acquaint you with my favorite brand of ground hot red pepper!
Y'know, I understand the "Great (Fire)Wall of China". Shouldn't this be the "Australian Great Barrier Firewall"? And - isn't that in danger of being destroyed by people poking and prodding at it, punching holes in it, etc.?
Windows Vista -> Windows 7 is another facelift. MicroSoft has been mostly honest about this. Yes, there have been conflicting rumors, some of which originated with MicroSoft personnel, but the source doesn't necessarily guarantee the accuracy of the information. MicroSoft's official line on Windows 7 is to quietly admit that it isn't a major code revision - in fact, I heard one rumor that they were planning to release it under the name "Windows 6.3" instead of the codename "Windows 7", to explicitly make it (sort of) clear that this is a minor revision, not a major one.
Not that I think Windows Vista is worth the DVD's it ships on, but the folks here might consider giving MicroSoft a break. My opinion of MicroSoft remains pretty low, but in this I think they're actually trying to do something right. They're not claiming to have invented a new paradigm or anything stupid like that, they're just saying that there's a new version of the Windows operating system coming soon.
That's right - let the Governments of the world fix the internet by legislation; after all, we all know how well the government understands the tubes of the intarweb. Perhaps Al Gore could be tapped to spearhead this incredibly important piece of legislation.
So . . . has any pulsar ever been gobbled up by a microscopic black hole, or turned into strange matter (stranglets), or caused a quantum vacuum explosion? No?
GO CERN!
candidate for use in an encryption scheme. Problems of class NP are especially useful in this area.
Now, if they'd said they were going to incorporate a Peltier device on the chip die to let 'em run heavily overclocked and ice-cold, I might've fallen for it . . .
Legal nonsense, but when coupled with an honest, good faith campaign to ensure that no one creates links to their website, well . . . when even the Goog can't display your links, what's the point of having a web presence?
I can also see this in court - "Your honor, we didn't reverse engineer functionality {blah} - our system is still in beta. We were planning to implement {blah} all along. We didn't steal the idea from so-and-so."
Do no evil? Hmmm . . . how 'bout "Free the beast". Has anybody here looked at the price of Google stock lately? I use their tools regularly (frankly, I'm glad as hell that Google exists), but I don't fool myself. They're definitely in it for the money. Then again, so am I.
It's okay - go back to your bananas and your blocks, leave everything up to those of us who can read. It's okay.
This legislation will force industry to develop and pay for it, regardless of whether the customers want it or not. Yes, we all want encryption on everything; but an overwhelming majority of computer users don't care enough to actually do anything, even though it would only take a bit of time and effort. Now, what happens when your bank send you your private encryption key and instructions? Most recipients will either delete or (at best) ignore the key. Later that month imagine their anger when their bank statement is encrypted and they have no idea how to decrypt it? Or do you really get the impression that the average American (Nevadan?) consumer is intelligent enough to implement, say, GPG? If so, do you think the average consumer is energetic enough to do so?
Leave this job up to market forces - the free-enterprise economy is infinitely more responsive to the needs and wants of the average consumer than is the Federal or even any of the State governments.
Yeesh! Who says they'll end up using a wavelength in the visible spectrum? Can you say "maser"?
Consider - if a bank sent new ATM cards with the pin in the same envelope as the card, most consumers would go immediately berserk. The institution in question would rapidly see an erosion of their customer base, as well as being found liable for any losses incurred by people who had their mail intercepted by thieves.
That same bank can blithely send out e-mails with user account names, numbers and passwords all in one convenient, easy-to-sniff package and nobody gets upset. How often has anybody here clicked on "forgot my username/password" only to get a nice, convenient clickable link which allows unfettered access to private, smooth, creamy soft personal information? The solution isn't for the government to legislate the use of encryption; rather, it's a matter for market pressure. Let enough people become unhappy over the cavelier treatment their personal information garners from a corporation - they'll vote with their wallets, if they are once educated regarding the situation.
That last phrase is the hard part though, isn't it?
How do they keep 'em patched, I wonder? Must be some kind of clustering scheme, since we all know that most Windows patches require rebooting, yes?
Yeah, give me enough hardware and I can make almost any OS (MS-DOS anyone) do whatever is needed in a multitasking server environment. Then again, Microsoft is quite famous for delivering software which requires (b)leeding edge hardware to run adequately.
"The price of Freedom is eternal vigilance." - who said that again?