The problem with wiring the accelerator into the CPU is that, although the CPU can perform the calculation faster, it does not actually free the CPU from having to do the packet processing. In addition to CPU time spent, you also need to consider interrupt overhead, which for high-speed networks (like 10GbE) is pretty significant. A separate TCP offload engine, with hardware encryption support, and access to memory via DMA, can significantly reduce the amount of time a CPU spends processing packets. It just interrupts the CPU when a decrypted TCP payload is ready and waiting in memory. And since your add-in card doesn't need a large instruction set, you can make it very, very fast.
I can vouch for 250 Vista licenses with XP downgrade rights right here. Also, the same number of Office 2007 with Office 2003 downgrade rights.
At home, my computer purchases don't happen at the level of "this is a whole, functioning computer", rather, I buy in parts, so I have no need for downgrade rights there. Of course, I don't run Windows either, but that's a different story.
That's a weird bit of logic. If people are running Firefox on 2000/XP, then they're not running IE6, by virtue of running Firefox. How will taking Firefox away from them drive them away from IE6? Are you thinking that they'll up and move themselves to Vista? Most Windows users I know would probably switch back to IE if they had to do that.
I think part of the problem is that ATM machines have, in the past, not used IP networks, because there was always a need to lay down a line (or a modem) that would connect to the financial network. Many financial networks predate the Internet, and many of them have stricter requirements than typical IP traffic (like QoS), and so, in many cases, you see other kinds of network architectures (like X.25). Given those conditions, strong encryption did not always make sense.
Now, there's nothing stopping you from using a higher-level protocol like SSL with other network architectures, but ATMs already have their own security mechanisms that predate SSL by a long shot, and the use of SSL, at least culturally, is tied pretty closely with TCP/IP. What surprises me, though, is that the HSMs must decrypt a message at every interchange, and re-encrypt it. I'm sure financial networks were around before asymmetric encryption was widely known or used, but they've had a long time to do this the right way now. The fact that these networks are still vulnerable to MITM attacks is pretty shocking.
Anyway, I don't know a whole lot about financial networks. Anyone care to fill us in?
I'm not crazy about the ad, either, but to play devil's advocate, this ad didn't exactly tell you what it was advertising either. On many levels, that ad should have been a flop (no information about the product, decidedly intellectual metaphor, etc) But you left the ad knowing that you wanted to know more. I think that's the approach the Linux folks are trying here.
In my opinion, that approach is fine, but this ad doesn't really make me want to find out more. Maybe they need to hire Ridley Scott.
Bush's motorcade was pelted with snowballs on the way to his inauguration while Obama got a party.
Maybe that's because many of us felt that presidents should not be chosen by the Supreme Court, especially when that choice runs contrary to the popular vote. Bush's re-election was more disturbing to me because he clearly had more Americans behind him (but why, I don't know).
Or maybe the causes of stranding are a many-to-one relation, i.e., that there is more than one cause, and that use of sonar is only one of them. E.g., you find dead birds with broken necks all the time. It is disingenuous to say that windows are the cause of all broken bird necks, and we can point out that people have found dead birds with broken necks even in antiquity. But it is equally disingenuous to say that windows have nothing to do with it.
Putting aside, for a moment, the fact that 1.5*2 is a lot closer to 3.35 than 1.5*3, a large proportion of that 3.35 Terawatts/year is not converted into heat. It is converted into mechanical work (machinery, etc) and other things (televisions, computers, communications and so on). Now, the net heat dissipation of the United States from electrical power-- I'm not sure anyone knows that, but by conservation of energy, we know it cannot be the total 3.35 Terawatts/year. Because some of it is being used for other things.
Which is almost beside the point. Let's say that that 3.35 Terawatts/year is converted into heat (i.e., it is used 100% inefficiently), and that that heat dissipation directly causes 2 additional hurricanes/year. Well, by virtue of the fact that we've removed 2 hurricanes/year's worth of energy from the wind by capturing it in a turbine, in order to power our 2 hurricanes/year, at worst, we are better off than we are in our actual state. There is no net gain in wind energy in the system. Right now, in our actual state, again assuming the worst case, 100% inefficiency, we are pulling 2 hurricanes worth of energy out of the ground and putting it straight into the atmosphere. That's a net gain of 2 hurricanes. By this logic, wind turbines are easily a benefit over conventional power production. Since we used the worst case above for wind production, and the reality of it is better than that, the argument for wind looks even better.
No shit. There's wind in lots of places. There are moving bodies of water in lots of places. There is sun in lots of places. There is geothermal energy in lots of places. You can take advantage of at least one of these things in many of the places we live.
I was driving across West Virginia a few weeks ago and I saw lots of wind farms. It was pretty cool. And West Virginia is nowhere near the ocean. And it was pretty f'ing windy, I might add, at least while I was there.
CO2 is a bit different because it's (according to the prevailing wisdom) screwing with one of the inputs
Technically, it's screwing with the outputs, since the effect of increased CO2 is to reduce the amount of radiative cooling. Lots of energy hits the Earth and bounces back into space. Consider the fact that we can directly pick out infrared emissions of exoplanets using telescopes. That's a lot of energy leaving the surface of a planet!
We know that energy must be leaving the system, because we're capturing it with a turbine. We know that the amount of energy we remove from the system (we can define the system here as "the Earth", I suppose) is [energy harnessed] + [conversion loss]. Conversion losses probably happen in the mechanical linkages, transmission losses on the power lines due to impedance, and so on. These things generate heat. I don't know what the net effect of reducing wind speed would be. Maybe it increases the local temperature because there is less convection, but the global temperature is decreased because there is less friction from air movement? I don't know a whole lot about fluid mechanics.
In the end, I think an engineer/scientist/physicist would ask: how significantly do these things impact their surroundings? On the global scale, it may be insigificant. On the local scale, it may also be insignificant-- the only way to find out is to try it (and then revise our model).
What we do know is that traditional (carbon-powered) power plants presently generate an enormous amount of heat, noise and emissions, since they rely on combustion. So even if wind generation does indeed have enough significance to alter the global climate, we may find that we still prefer it to our present technology.
I welcome comments from anyone who has actual figures they can post.
"Ass-clown" is a colloquial expression for "idiot [in the extreme]", but I think the genesis of the word was originally more along the lines of "queer" or "fag". Some of my friends use it here in New England. Let's just say it is not generally used in polite company.
Of course, the problem with nitrogen is that nitrogen = finer bubbles in the head = waiting longer to drink your beer. Actually, I was in a pub in Ireland once, and an (American, strike 1) girl ordered a half-pint of Guinness (which is strike 2 against her in the barkeep's mind; a half pint?). When he finished pouring it, he set it on the counter, and she immediately reached for it (strike 3). He smacked her hand, and told her to sit down and wait. She had no idea what was going on.
She later tried to order a "Smith Wicks" and was told, "never heard of it"...
You can deny it all you want, but it doesn't change the fact that marriage (among people with with European ancestry, anyway) was essentially a financial transaction for millennia. That didn't start to change until the mid 1500's, when the Catholic church stepped in and demanded that marriages be approved by priests.
That's a good question. Unfortunately the answer is expensive to determine. At the small-to-midsize operation that I run, we don't have enough stuff to give us useful data. We've favored the cheaper solution, where applicable (i.e., SATA disks), unless they turned out to be horrible (which has never happened). Not to mention, as long as you're fairly attentive, and replace disks when they need replacing, it doesn't much matter if SATA disks fail more often than SCSI. If we were running thousands of disks, though, it might be a different story.
People continue to run equipment after the manufacturer stops making it, you know. Those Xserve RAIDs are still in service, even though the Xserves themselves are not. They are probably the one somewhat decent Apple product; we have Infortrend RAIDs that are much better, though. They're part of a StorNext network now.
As for the Xserves, we used them for spare parts until they were basically devoid of anything useful. They reached a threshhold where they were no longer economical to support and repair. My problem with Apple's enterprise stuff is that this threshold is pretty low. Compared to, say, Supermicro, where almost all of the parts are either off-the-shelf or easily made (e.g., front panel ribbon cable), the useful lifetime for an Apple enterprise product is extremely low, because all the equipment comes at a premium. For us, we could only continue to justify that expense so long as we were using Apple software.
Since we now run Netatalk on Linux (integrated with AD, with much less effort than the comparable Apple product, I might add), and our Mac clients use OSS or Microsoft services, we have no need to Apple software. We'd sell off the old Xserve cadavers to eBay, but corporate policy forbids it. So they sit there collecting dust.
Lots of useful equipment gets discarded in enterprise IT because it is no longer economical to use. Do you actually have an enterprise IT experience, or are you just talking out of your ass?
My brother and I have a working theory that your typical male, given a choice between sex and probable death, would definitely choose the sex. E.g., while engaging in some typical male activity (making something explode, consuming vast quantities of alcohol, traveling way to fast; take your pick), you become mortally wounded. However, since clearly what I am describing here is a Bruce Willis movie, you are shortly thereafter presented with the opportunity of engaging in sex with a beautiful woman (or, you know, a man, if that's more your style). Do you take it? We're fairly certain that a large percentage of the male population would, and that, in fact, this is the quintessential test of manliness.
Yeah, but the kicker is, sometimes the support still sucks. E.g., we have (er, HAD; Apple pissed me off to the point that I literally have a stack of Xserves that can turn to dust for all I care) a bunch of Xserves and Xserve RAIDs that we used for various tasks, mostly Mac-related.
One day, the RAID Admin tool notifies me via email that I have a failing disk in an array. OK, no problem, call Apple. We're paying for support, you know.
The guy on the phone was friendly, and said the disk was on its way. The next morning, I have a package waiting for me when I get in. It's a drive module, yay! I go into the server room, yank the failed module (conveniently designated by a red light) and insert the new one. RAID Admin proceeds to tell me that there was an unexpected error. Huh? I try again. Same thing. No additional information, just "unexpected error".
I call Apple back, and explain the problem, they run me through some GUI diagnostics, and in the end, cannot solve the problem. I tell them to hold off on sending me another disk just yet.
After trawling the Apple forums, I find out, hey, I can get all of the diagnostics and logs from the CLI, too. They're way more verbose. Verdict: Apple sent me a disk of the wrong capacity-- the ones in my RAID are 74.5 GB and this one is 73.x or something. Of course, they're _supposed_ to know exactly what kind of disks I have; that's the whole point of the service. Anyway, it eventually gets sorted out, but after the RAID sat there, operating without a hot spare for about a week.
Now, if this were the end of it, I would be forgiving of Apple. But I've had other problems. We had a CD reader fail in an Xserve. The Apple on-site person came out for this one. When he left (without checking in with me, of course), it still DID NOT WORK. This is despite the fact that I set him up with a workbench, full complement of tools, power, keyboard, and monitor to test with. The problem? He never bothered to plug the new CD drive into the machine. This is shoddy service. MINIMALLY, you test the part you just replaced, right?
But the icing on the cake was when a controller module in our Xserve RAID failed. I call Apple and they overnight a part. When I open the box, I have... a CD-ROM drive? I call Apple and say, hey, you sent me the wrong part. The support guy went so far as to call me a liar on the phone. He said that such a mix up was "impossible". He was convinced that I was going to return the box with a CD-ROM drive in it, and keep the shiny new controller that they sent me. It wasn't until I faxed them photocopies of the accompanying paperwork that they would believe me that it was their error, and even then, they CHARGED me for a SECOND part! It took our A/P department about a month to sort that out.
So fuck you very much Apple. Fucking rot in hell.
(And yes, I typed this message on an iMac. I like punishment; what can I say?)
Slashdot Mirror
All I want to know is if floppy computers can have hard disks. Or if we have to go back to floppy disks.
The problem with wiring the accelerator into the CPU is that, although the CPU can perform the calculation faster, it does not actually free the CPU from having to do the packet processing. In addition to CPU time spent, you also need to consider interrupt overhead, which for high-speed networks (like 10GbE) is pretty significant. A separate TCP offload engine, with hardware encryption support, and access to memory via DMA, can significantly reduce the amount of time a CPU spends processing packets. It just interrupts the CPU when a decrypted TCP payload is ready and waiting in memory. And since your add-in card doesn't need a large instruction set, you can make it very, very fast.
I can vouch for 250 Vista licenses with XP downgrade rights right here. Also, the same number of Office 2007 with Office 2003 downgrade rights.
At home, my computer purchases don't happen at the level of "this is a whole, functioning computer", rather, I buy in parts, so I have no need for downgrade rights there. Of course, I don't run Windows either, but that's a different story.
That's a weird bit of logic. If people are running Firefox on 2000/XP, then they're not running IE6, by virtue of running Firefox. How will taking Firefox away from them drive them away from IE6? Are you thinking that they'll up and move themselves to Vista? Most Windows users I know would probably switch back to IE if they had to do that.
I think part of the problem is that ATM machines have, in the past, not used IP networks, because there was always a need to lay down a line (or a modem) that would connect to the financial network. Many financial networks predate the Internet, and many of them have stricter requirements than typical IP traffic (like QoS), and so, in many cases, you see other kinds of network architectures (like X.25). Given those conditions, strong encryption did not always make sense.
Now, there's nothing stopping you from using a higher-level protocol like SSL with other network architectures, but ATMs already have their own security mechanisms that predate SSL by a long shot, and the use of SSL, at least culturally, is tied pretty closely with TCP/IP. What surprises me, though, is that the HSMs must decrypt a message at every interchange, and re-encrypt it. I'm sure financial networks were around before asymmetric encryption was widely known or used, but they've had a long time to do this the right way now. The fact that these networks are still vulnerable to MITM attacks is pretty shocking.
Anyway, I don't know a whole lot about financial networks. Anyone care to fill us in?
How about "no viruses"? I think most people would "get" that.
I'm not crazy about the ad, either, but to play devil's advocate, this ad didn't exactly tell you what it was advertising either. On many levels, that ad should have been a flop (no information about the product, decidedly intellectual metaphor, etc) But you left the ad knowing that you wanted to know more. I think that's the approach the Linux folks are trying here.
In my opinion, that approach is fine, but this ad doesn't really make me want to find out more. Maybe they need to hire Ridley Scott.
"People. What a bunch of bastards." -- Roy
Bush's motorcade was pelted with snowballs on the way to his inauguration while Obama got a party.
Maybe that's because many of us felt that presidents should not be chosen by the Supreme Court, especially when that choice runs contrary to the popular vote. Bush's re-election was more disturbing to me because he clearly had more Americans behind him (but why, I don't know).
Actually, I think we finally get to find out why Spock is so dern smart. It involves cutting people's heads open!
The really important question is whether the Enterprise's license plates say H3R035 on it, or something.
Or maybe the causes of stranding are a many-to-one relation, i.e., that there is more than one cause, and that use of sonar is only one of them. E.g., you find dead birds with broken necks all the time. It is disingenuous to say that windows are the cause of all broken bird necks, and we can point out that people have found dead birds with broken necks even in antiquity. But it is equally disingenuous to say that windows have nothing to do with it.
Putting aside, for a moment, the fact that 1.5*2 is a lot closer to 3.35 than 1.5*3, a large proportion of that 3.35 Terawatts/year is not converted into heat. It is converted into mechanical work (machinery, etc) and other things (televisions, computers, communications and so on). Now, the net heat dissipation of the United States from electrical power-- I'm not sure anyone knows that, but by conservation of energy, we know it cannot be the total 3.35 Terawatts/year. Because some of it is being used for other things.
Which is almost beside the point. Let's say that that 3.35 Terawatts/year is converted into heat (i.e., it is used 100% inefficiently), and that that heat dissipation directly causes 2 additional hurricanes/year. Well, by virtue of the fact that we've removed 2 hurricanes/year's worth of energy from the wind by capturing it in a turbine, in order to power our 2 hurricanes/year, at worst, we are better off than we are in our actual state. There is no net gain in wind energy in the system. Right now, in our actual state, again assuming the worst case, 100% inefficiency, we are pulling 2 hurricanes worth of energy out of the ground and putting it straight into the atmosphere. That's a net gain of 2 hurricanes. By this logic, wind turbines are easily a benefit over conventional power production. Since we used the worst case above for wind production, and the reality of it is better than that, the argument for wind looks even better.
No shit. There's wind in lots of places. There are moving bodies of water in lots of places. There is sun in lots of places. There is geothermal energy in lots of places. You can take advantage of at least one of these things in many of the places we live.
I was driving across West Virginia a few weeks ago and I saw lots of wind farms. It was pretty cool. And West Virginia is nowhere near the ocean. And it was pretty f'ing windy, I might add, at least while I was there.
CO2 is a bit different because it's (according to the prevailing wisdom) screwing with one of the inputs
Technically, it's screwing with the outputs, since the effect of increased CO2 is to reduce the amount of radiative cooling. Lots of energy hits the Earth and bounces back into space. Consider the fact that we can directly pick out infrared emissions of exoplanets using telescopes. That's a lot of energy leaving the surface of a planet!
We know that energy must be leaving the system, because we're capturing it with a turbine. We know that the amount of energy we remove from the system (we can define the system here as "the Earth", I suppose) is [energy harnessed] + [conversion loss]. Conversion losses probably happen in the mechanical linkages, transmission losses on the power lines due to impedance, and so on. These things generate heat. I don't know what the net effect of reducing wind speed would be. Maybe it increases the local temperature because there is less convection, but the global temperature is decreased because there is less friction from air movement? I don't know a whole lot about fluid mechanics.
In the end, I think an engineer/scientist/physicist would ask: how significantly do these things impact their surroundings? On the global scale, it may be insigificant. On the local scale, it may also be insignificant-- the only way to find out is to try it (and then revise our model).
What we do know is that traditional (carbon-powered) power plants presently generate an enormous amount of heat, noise and emissions, since they rely on combustion. So even if wind generation does indeed have enough significance to alter the global climate, we may find that we still prefer it to our present technology.
I welcome comments from anyone who has actual figures they can post.
Yes, a killing machine that runs on love. Good idea.
"Ass-clown" is a colloquial expression for "idiot [in the extreme]", but I think the genesis of the word was originally more along the lines of "queer" or "fag". Some of my friends use it here in New England. Let's just say it is not generally used in polite company.
Of course, the problem with nitrogen is that nitrogen = finer bubbles in the head = waiting longer to drink your beer. Actually, I was in a pub in Ireland once, and an (American, strike 1) girl ordered a half-pint of Guinness (which is strike 2 against her in the barkeep's mind; a half pint?). When he finished pouring it, he set it on the counter, and she immediately reached for it (strike 3). He smacked her hand, and told her to sit down and wait. She had no idea what was going on.
She later tried to order a "Smith Wicks" and was told, "never heard of it"...
You can deny it all you want, but it doesn't change the fact that marriage (among people with with European ancestry, anyway) was essentially a financial transaction for millennia. That didn't start to change until the mid 1500's, when the Catholic church stepped in and demanded that marriages be approved by priests.
My girlfriend would double-over laughing. She enjoys a quick wit and sense of humor.
Sorry, this just reminded me of...
"You know, when you, like, you grab a woman's breast and it's... And you feel it and... it feels like a bag of sand when you're touching it."
"Bag of sand?"
"You know what I mean. Why don't we just play? Why don't you just deal the cards?"
"What are you talking about?"
"Have you ever felt a breast before, man?"
"Yes."
"Dude, are you gay?"
"No, I'm not gay. I've been with tons of women."
"I touched a guy's balls at Hebrew school once."
"Dude, it's not a big deal. You like to fuck guys."
"I'm cool. I got friends who fuck guys, in jail."
"No, I'm not gay. No, I've borked a lot of women in my day."
"You've "borked"?"
"Hold up. Yo, answer this question: Are you a virgin?"
"Are you a virgin?"
"Yeah, not since I was 10."
"It all makes sense. You're a virgin."
Apropos.
That's a good question. Unfortunately the answer is expensive to determine. At the small-to-midsize operation that I run, we don't have enough stuff to give us useful data. We've favored the cheaper solution, where applicable (i.e., SATA disks), unless they turned out to be horrible (which has never happened). Not to mention, as long as you're fairly attentive, and replace disks when they need replacing, it doesn't much matter if SATA disks fail more often than SCSI. If we were running thousands of disks, though, it might be a different story.
People continue to run equipment after the manufacturer stops making it, you know. Those Xserve RAIDs are still in service, even though the Xserves themselves are not. They are probably the one somewhat decent Apple product; we have Infortrend RAIDs that are much better, though. They're part of a StorNext network now.
As for the Xserves, we used them for spare parts until they were basically devoid of anything useful. They reached a threshhold where they were no longer economical to support and repair. My problem with Apple's enterprise stuff is that this threshold is pretty low. Compared to, say, Supermicro, where almost all of the parts are either off-the-shelf or easily made (e.g., front panel ribbon cable), the useful lifetime for an Apple enterprise product is extremely low, because all the equipment comes at a premium. For us, we could only continue to justify that expense so long as we were using Apple software.
Since we now run Netatalk on Linux (integrated with AD, with much less effort than the comparable Apple product, I might add), and our Mac clients use OSS or Microsoft services, we have no need to Apple software. We'd sell off the old Xserve cadavers to eBay, but corporate policy forbids it. So they sit there collecting dust.
Lots of useful equipment gets discarded in enterprise IT because it is no longer economical to use. Do you actually have an enterprise IT experience, or are you just talking out of your ass?
My brother and I have a working theory that your typical male, given a choice between sex and probable death, would definitely choose the sex. E.g., while engaging in some typical male activity (making something explode, consuming vast quantities of alcohol, traveling way to fast; take your pick), you become mortally wounded. However, since clearly what I am describing here is a Bruce Willis movie, you are shortly thereafter presented with the opportunity of engaging in sex with a beautiful woman (or, you know, a man, if that's more your style). Do you take it? We're fairly certain that a large percentage of the male population would, and that, in fact, this is the quintessential test of manliness.
Yeah, but the kicker is, sometimes the support still sucks. E.g., we have (er, HAD; Apple pissed me off to the point that I literally have a stack of Xserves that can turn to dust for all I care) a bunch of Xserves and Xserve RAIDs that we used for various tasks, mostly Mac-related.
One day, the RAID Admin tool notifies me via email that I have a failing disk in an array. OK, no problem, call Apple. We're paying for support, you know.
The guy on the phone was friendly, and said the disk was on its way. The next morning, I have a package waiting for me when I get in. It's a drive module, yay! I go into the server room, yank the failed module (conveniently designated by a red light) and insert the new one. RAID Admin proceeds to tell me that there was an unexpected error. Huh? I try again. Same thing. No additional information, just "unexpected error".
I call Apple back, and explain the problem, they run me through some GUI diagnostics, and in the end, cannot solve the problem. I tell them to hold off on sending me another disk just yet.
After trawling the Apple forums, I find out, hey, I can get all of the diagnostics and logs from the CLI, too. They're way more verbose. Verdict: Apple sent me a disk of the wrong capacity-- the ones in my RAID are 74.5 GB and this one is 73.x or something. Of course, they're _supposed_ to know exactly what kind of disks I have; that's the whole point of the service. Anyway, it eventually gets sorted out, but after the RAID sat there, operating without a hot spare for about a week.
Now, if this were the end of it, I would be forgiving of Apple. But I've had other problems. We had a CD reader fail in an Xserve. The Apple on-site person came out for this one. When he left (without checking in with me, of course), it still DID NOT WORK. This is despite the fact that I set him up with a workbench, full complement of tools, power, keyboard, and monitor to test with. The problem? He never bothered to plug the new CD drive into the machine. This is shoddy service. MINIMALLY, you test the part you just replaced, right?
But the icing on the cake was when a controller module in our Xserve RAID failed. I call Apple and they overnight a part. When I open the box, I have... a CD-ROM drive? I call Apple and say, hey, you sent me the wrong part. The support guy went so far as to call me a liar on the phone. He said that such a mix up was "impossible". He was convinced that I was going to return the box with a CD-ROM drive in it, and keep the shiny new controller that they sent me. It wasn't until I faxed them photocopies of the accompanying paperwork that they would believe me that it was their error, and even then, they CHARGED me for a SECOND part! It took our A/P department about a month to sort that out.
So fuck you very much Apple. Fucking rot in hell.
(And yes, I typed this message on an iMac. I like punishment; what can I say?)