But it still goes boom right? Err wait no sound... well actually... no.. but there would be a pressure wave... no air to carry it... DAMN YOU SCI-FI CHANNEL!!
Dear God if they make batteries like they made cars you'll get a nice looking battery with lousy battery life and replacing a screw on the battery will run you $40 buck but you'll have to take it to the dealer to reset the Op Codes indicating you replaced the screw...
I can see it now, swap a batttery? Sorry $50 buck to plug in the Bear unit to reset the battery ID. No you can't take it to your local garage, battery codes are dealer only....
I'm mean seriously, a star exploding is a star exploding. Mario or Super Mario. He's still a fat plumber who eats shrooms...
I bet if the highly paid scientists found it they'd be touting the "Strongest NOVA ever see discovered" where as some plucky kid finds it they're like "umm weakest Super nova ever...."
Word play is fun...
It is almost like asking "Is it an A- or a B+" or the musical types the whole sharp flat deal...
I cannot in my experience accept theft for food at least in the Twin Cities MN. Nearly 3/4th of all convictions for theft included drug possession, not food possession.
By looking at root causes in crime it can help policy makers see where the best return on money can be spent. More cops doesn't mean less crime always. If it was easier for you to get out of the ghetto would you agree that the need for crime to get out of the ghetto decrease?
The problem is a bit more complex then CFS\FGS\etc.. It is a weakness in x86 in reality that would take exotic workarounds.
A VM in general is a single process. Process Time allocated isn't neccessarily reflective of performance especially when time sensitive real time transactions are required INSIDE the VM. Halt or suspend a process too long and the task inside the VM might crap out.
Part of the problem is having hard and soft limits on processing load and controlling which other systems can borrow from.
Using your example (a very good one at that) the best an admin can do is have a GROUP called say VM and within use nice values inside. But with SLAs you have to have a hard floor for a VM, used or otherwise as well has a hard ceiling.
Here is an example that you can try to work out:
VM1 must have 25% access to a processor at all times and can use up to 95% of the processor.
VM2 Must have access to 25% of the processor at all times that VM1 is not using. And can use as much as it can use as needed any unused CPU usage.
VM2 Must have access to 10% of the processor at all time that VM1 is not using but can never use more then 15% of the processor regardless of what VM1 or VM2 is using.
VM3 Must have access to 20% of the processor at all times that VM1 is not using but can never use more then 28% of the processor.
At no time can any VM get less then 1% of the CPU time.
Now try setting that up using any existing scheduling mechanics in Linux. (I just threw that together as an example real quick.)
Those are the kind of questions that SLA based provisioning has to deal with.
The tricky part is the hard limits. Once a process is kicked off many times it becomes difficult, if not impossible to throttle it (rather then say, suspend it for a while, a.k.a put a job on hold) as it executes based on an SLA with hard and soft limits.
See nice values are not % weights, they are priority indicators nor can you get a MIPS weighting to allocate either. Run time dynamic systems exist that will re-nice things to try and meet a target but those are not fine grained enough to write contracts against.
e.g.
VM1 Running and tracking it's CPU time is at 25% VM2 suddenly exceeds it's allocation so the tracker renices it to drop it's priority until it goes back into it's sla.
Awww hell this will take too long: short version: User space tools suck at managing this.
The point of VM tech is isolation and why mainframes are still popular. My experience so far is most is code migration rather then VM images. I've seen a few DEV VMs try to capitalize PROD CPU but they are automatically live migrated off when they exceed 60% utilization.
The method for handling infrastructure changes from company to company. The leading trend I see currently is the "Fabric" architecture where you have large pools of hardware akin to mainframes. There really isn't new infrastructure in the existing pools. New infrastructure is a whole new fabric.
FAB1 FAB2 FAB3
If they deploy a new FAB4 pool then the test VMs are transfered first then eventually PROD VMS. The older FABS are either decomissioned or turned into LOWPERF, MEDPERF pools.
The trend I see is high mobility in VM technology. Not saying you are incorrect but I am retired from system administration an enginnering (I am now a humble Business Analyst... ah peace..)
I know that security concerns for data protection prevent production and test sharing a SAN system but so far I have not seen any vulnerabilities in one VM leaking data to another such that GLBA\SOX\CISP would prohibit VMs sharing Iron (data is the key, not execution.) But that may change.
It will be interesting to see how this plays out in the next 5 years.
The point is with over 1 billion stars in our own GALAXY I find the need to look in a neighboring galaxy moot. We can look at well over a billion stars in this galaxy to study our own solar system. Why the next galaxy?
You say that like it was a good thing? There were a lot of people that left because there were a lot of people playing that "didn't get it". It's like those poor souls on RPG servers for WoW. They want to role play but there are just too many people that "Don't get it."
It depends on which crowd you were in. The early adopters and insiders really didn't like it when every day joe started playing.
Yeah but some shops just won't do business in non-Intel. And those wanting to research clustered VM fabrics for managing VMs POWER5/6 isn't an option. Just exploring Linux\KVM's possibilities.
I live in Minnesota, I am aware. But brilliant and intelligent doesn't mean they will overcome inexperience either.
Unlike "hard sciences" you are dealing in the realm of public opinion, interpretation, and paperwork so convoluted and twisted a single mis-written sentence can result in a thrown out case.
Einstein's work stood on it's own. Proofs and equations can be tested regardless if you like communists or not. Regardless if you liked jews or not. That is the one thing I admire most about scientists, you are hard pressed to argue with a sound equation regardless of personal opinion.
In the legal world influence a judge has on the outcome of a case, no matter how sound the underlying case, can greatly influence the outcome. You piss the judge off your more then likely going to lose. Experience tells you who your dealing with, the nuances of the court itself, and what pitfalls to watch out for. How to work a jury, how to generate symapthy.
Just voice my concerns is all. And as far as responding to you, I always try, your opinion is always worth my time. (sans my every day life stuff that is > Internet time.:) )
I never said opposition is not an option, but corrdinated opposition is much more effective then a few lone wolves going off.
You may win the battle but lose the war. I'd rather have a comprehensive plan to win the war rather then carve out a piecemeal strategy.
All I have stated is this is far beyond the risk threshold I am comfortable with.
That is the problem with the Internet, some people just can't comprehend an opinion anymore. Many have just been brainwashed I suppose to assume that an opinion is some kind of edict or decree.
"I don't like bean sprouts in fried rice."
Who in their right mind would assume that I am some how implying we should ban them, or there should be no bean sprouts in fried rice? Yes it is that silly people... How many posts make that assumption? I swear there must be a part of the human brain that stopped working as a whole.
Relax... Breath... and stop drawing spurious conclusions. I know school has taught you that criticism is all forms is bad but simply put, I'm just voicing my concerns. Chill.
If you are not worried NYCL I sleep all that much more easy but I have seem in my life far too many "hot shots" of any field, even when paired with experts, screw up. I don't like the path this could take is all. I mean just looking when the wife was busting through her studies at Hamline I swear it gives me the chills on how appeals move around.
As far as attitude goes, how many judges have you found that appreciate "attitude"? That's what worries me.
And for your a-e I'd like to ask about (f) Disciplined? e.g. "they're not going to end up mouthing off to the judge"
The verdict gives a path to appeal. The higher you go the more politically entrenched the judge. The greater the risk of a pro-RIAA verdict. Plain and simple.
"When I was young science focused on the betterment of human life on Earth, not wasting tax payer money staring at galaxies far far away.":)
The whole point is we have a whole Milky Way to work with, putzing around Andromeda and other galaxies seems pretty low on the totem pole when public money is involved.
Well I can see practical reasons for studying Saturn and Jupiter for climate modelling and sub-nuclear gas giants. But we also can get there and back in a lifetime. Studying gas giants 500 light years away I see little value beyond "Science for Science's sake."
IMHO I would have to say yes. With the rise of multi-core systems I found that FORTRAN has some programming habits that would carry into multi-threaded and multi-core programming. While I don't think it has a high relevancy for finding employment I do think that FORTRAN, along with at least a 3 credit hour course in assembly is crucial for fundamental programming skills.
I think we can drop PASCAL (If they haven't already) and swap it with either Ruby or PERL for undergraduates.
Slashdot Mirror
I smell European insidfe joke here... explain...
WHat no Guess Who jokes?
But it still goes boom right? Err wait no sound... well actually... no.. but there would be a pressure wave... no air to carry it... DAMN YOU SCI-FI CHANNEL!!
I honestly don't know. Some days I am just in the mood to burn karma and I fail miserably. I was shooting for Funny actually.
Dear God if they make batteries like they made cars you'll get a nice looking battery with lousy battery life and replacing a screw on the battery will run you $40 buck but you'll have to take it to the dealer to reset the Op Codes indicating you replaced the screw...
I can see it now, swap a batttery? Sorry $50 buck to plug in the Bear unit to reset the battery ID. No you can't take it to your local garage, battery codes are dealer only....
Pet Peeve Alert:
Weakest Supernova or STRONGEST NOVA?
I'm mean seriously, a star exploding is a star exploding. Mario or Super Mario. He's still a fat plumber who eats shrooms...
I bet if the highly paid scientists found it they'd be touting the "Strongest NOVA ever see discovered" where as some plucky kid finds it they're like "umm weakest Super nova ever...."
Word play is fun...
It is almost like asking "Is it an A- or a B+" or the musical types the whole sharp flat deal...
I cannot in my experience accept theft for food at least in the Twin Cities MN. Nearly 3/4th of all convictions for theft included drug possession, not food possession.
By looking at root causes in crime it can help policy makers see where the best return on money can be spent. More cops doesn't mean less crime always. If it was easier for you to get out of the ghetto would you agree that the need for crime to get out of the ghetto decrease?
The problem is a bit more complex then CFS\FGS\etc.. It is a weakness in x86 in reality that would take exotic workarounds.
A VM in general is a single process. Process Time allocated isn't neccessarily reflective of performance especially when time sensitive real time transactions are required INSIDE the VM. Halt or suspend a process too long and the task inside the VM might crap out.
Part of the problem is having hard and soft limits on processing load and controlling which other systems can borrow from.
Using your example (a very good one at that) the best an admin can do is have a GROUP called say VM and within use nice values inside. But with SLAs you have to have a hard floor for a VM, used or otherwise as well has a hard ceiling.
Here is an example that you can try to work out:
VM1 must have 25% access to a processor at all times and can use up to 95% of the processor.
VM2 Must have access to 25% of the processor at all times that VM1 is not using. And can use as much as it can use as needed any unused CPU usage.
VM2 Must have access to 10% of the processor at all time that VM1 is not using but can never use more then 15% of the processor regardless of what VM1 or VM2 is using.
VM3 Must have access to 20% of the processor at all times that VM1 is not using but can never use more then 28% of the processor.
At no time can any VM get less then 1% of the CPU time.
Now try setting that up using any existing scheduling mechanics in Linux. (I just threw that together as an example real quick.)
Those are the kind of questions that SLA based provisioning has to deal with.
The tricky part is the hard limits. Once a process is kicked off many times it becomes difficult, if not impossible to throttle it (rather then say, suspend it for a while, a.k.a put a job on hold) as it executes based on an SLA with hard and soft limits.
See nice values are not % weights, they are priority indicators nor can you get a MIPS weighting to allocate either. Run time dynamic systems exist that will re-nice things to try and meet a target but those are not fine grained enough to write contracts against.
e.g.
VM1 Running and tracking it's CPU time is at 25%
VM2 suddenly exceeds it's allocation so the tracker renices it to drop it's priority until it goes back into it's sla.
Awww hell this will take too long: short version: User space tools suck at managing this.
Hence why I ask, if no one asked they'd have no reason to progress that way.
The point of VM tech is isolation and why mainframes are still popular. My experience so far is most is code migration rather then VM images. I've seen a few DEV VMs try to capitalize PROD CPU but they are automatically live migrated off when they exceed 60% utilization.
The method for handling infrastructure changes from company to company. The leading trend I see currently is the "Fabric" architecture where you have large pools of hardware akin to mainframes. There really isn't new infrastructure in the existing pools. New infrastructure is a whole new fabric.
FAB1
FAB2
FAB3
If they deploy a new FAB4 pool then the test VMs are transfered first then eventually PROD VMS. The older FABS are either decomissioned or turned into LOWPERF, MEDPERF pools.
The trend I see is high mobility in VM technology. Not saying you are incorrect but I am retired from system administration an enginnering (I am now a humble Business Analyst... ah peace..)
I know that security concerns for data protection prevent production and test sharing a SAN system but so far I have not seen any vulnerabilities in one VM leaking data to another such that GLBA\SOX\CISP would prohibit VMs sharing Iron (data is the key, not execution.) But that may change.
It will be interesting to see how this plays out in the next 5 years.
The point is with over 1 billion stars in our own GALAXY I find the need to look in a neighboring galaxy moot. We can look at well over a billion stars in this galaxy to study our own solar system. Why the next galaxy?
drat the political science major has outed me! (hides) Correct.
You say that like it was a good thing? There were a lot of people that left because there were a lot of people playing that "didn't get it". It's like those poor souls on RPG servers for WoW. They want to role play but there are just too many people that "Don't get it."
It depends on which crowd you were in. The early adopters and insiders really didn't like it when every day joe started playing.
Yeah but some shops just won't do business in non-Intel. And those wanting to research clustered VM fabrics for managing VMs POWER5/6 isn't an option. Just exploring Linux\KVM's possibilities.
Out of millions of cases you came up with 3? Epp! :) And of the 3 they were made into movies? Double Epp!
I live in Minnesota, I am aware. But brilliant and intelligent doesn't mean they will overcome inexperience either.
Unlike "hard sciences" you are dealing in the realm of public opinion, interpretation, and paperwork so convoluted and twisted a single mis-written sentence can result in a thrown out case.
Einstein's work stood on it's own. Proofs and equations can be tested regardless if you like communists or not. Regardless if you liked jews or not. That is the one thing I admire most about scientists, you are hard pressed to argue with a sound equation regardless of personal opinion.
In the legal world influence a judge has on the outcome of a case, no matter how sound the underlying case, can greatly influence the outcome. You piss the judge off your more then likely going to lose. Experience tells you who your dealing with, the nuances of the court itself, and what pitfalls to watch out for. How to work a jury, how to generate symapthy.
Just voice my concerns is all. And as far as responding to you, I always try, your opinion is always worth my time. (sans my every day life stuff that is > Internet time. :) )
Nahh I'm not a legal expert to make such assertions on precident, I am referring to the process as a whole in abstract.
I never said opposition is not an option, but corrdinated opposition is much more effective then a few lone wolves going off.
You may win the battle but lose the war. I'd rather have a comprehensive plan to win the war rather then carve out a piecemeal strategy.
All I have stated is this is far beyond the risk threshold I am comfortable with.
That is the problem with the Internet, some people just can't comprehend an opinion anymore. Many have just been brainwashed I suppose to assume that an opinion is some kind of edict or decree.
"I don't like bean sprouts in fried rice."
Who in their right mind would assume that I am some how implying we should ban them, or there should be no bean sprouts in fried rice? Yes it is that silly people... How many posts make that assumption? I swear there must be a part of the human brain that stopped working as a whole.
Relax... Breath... and stop drawing spurious conclusions. I know school has taught you that criticism is all forms is bad but simply put, I'm just voicing my concerns. Chill.
If you are not worried NYCL I sleep all that much more easy but I have seem in my life far too many "hot shots" of any field, even when paired with experts, screw up. I don't like the path this could take is all. I mean just looking when the wife was busting through her studies at Hamline I swear it gives me the chills on how appeals move around.
As far as attitude goes, how many judges have you found that appreciate "attitude"? That's what worries me.
And for your a-e I'd like to ask about (f) Disciplined? e.g. "they're not going to end up mouthing off to the judge"
The verdict gives a path to appeal. The higher you go the more politically entrenched the judge. The greater the risk of a pro-RIAA verdict. Plain and simple.
2 + -1 = 1
"When I was young science focused on the betterment of human life on Earth, not wasting tax payer money staring at galaxies far far away." :)
The whole point is we have a whole Milky Way to work with, putzing around Andromeda and other galaxies seems pretty low on the totem pole when public money is involved.
Well I can see practical reasons for studying Saturn and Jupiter for climate modelling and sub-nuclear gas giants. But we also can get there and back in a lifetime. Studying gas giants 500 light years away I see little value beyond "Science for Science's sake."
IMHO I would have to say yes. With the rise of multi-core systems I found that FORTRAN has some programming habits that would carry into multi-threaded and multi-core programming. While I don't think it has a high relevancy for finding employment I do think that FORTRAN, along with at least a 3 credit hour course in assembly is crucial for fundamental programming skills.
I think we can drop PASCAL (If they haven't already) and swap it with either Ruby or PERL for undergraduates.
I get it but you could just as easily validate that looking at the far end of the Milky Way too.