It would be way better to do this as a libc pre-load shim. just overload "open" to get the path names/ip addresses of what file/connection is associated with each fd, then override read and write to track i/o rates, and block when exceeded. Could also
do it based on iops...(counting each read/write as an op, and giving op budget...)
Have a little config file consulted by the shim library:
I dunno what your use case is, you could use this for stuff you start as a user, or added
to the system config, and have it apply to all users. To do system-wide stuff, tc is already
plenty good enough, thought it only applies to network stuff. Dunno of anything for limiting b/w to local disk.
anyway, you don't need any kernel anything for this feature. basic rule is if it can be done in user space, it very often ought to be done there.
I call B.S. on the paranoid bit, they are simply being realistic in the strange world of business IT.
Crappy systems acquired incrementally over a decade by random contractors was tested for acceptance, piecemeal. Every one of them built for IE6 and only IE6. Those contractors are long gone. Anyone who knows anything about the applications, is likely also long gone, did they keep source code? not likely. Now to upgrade, they have to go back over a decade's worth of deployments and re-test and re-certify. That's expensive. And if the people on staff are just those who run recipes and tell people how to hold the mouse, that means hiring consultants to do the re-certifications. so are you willing to spend a few billion, upgrading from IE6 to IE8? What is the benefit to tax payers? pretty minimal imo.
The problem is that the "IT Architecture" used by most organisations is "something that works on what we have now" without the slightest appreciation of how ephemeral "now" is. Most organizations don't give a rats behind for open standards, and the result is the common rats nest, which is paradise for contractors and vendors. It's lock-in through entanglement.
Those people who refused are very rational, and very trapped. Their problem is the result of rampant ignorance and very poor decision making, but it is typical of many large organizations.
I hope that RJ-45 does become the AV standard that takes over, and that it is a real networking standard (not some bastardized CE/AV limited weird re-invented hodge podge that just uses the same connectors) I think you could take the idea further, and have equipment with just multiple RJ-45's/Cat6, and have the equipment do bonding/trunking. If you don't have a 4K TV, maybe one cat6 cable will be enough. Get a 4D Blue ray? add some patch cables of the same cheap variety, and you are good to go.
But on the other hand...All video signals can/are compressed. The over the air standard is MPEG2 compressed 1080i which fits in 6 MHZ of bandwidth. as in about 6 mbits... as in 1/200th of 10 Gbit/s. All TV's have MPEG decoders because, well, that's what they need to be TV's. Just standardizing compression over tcp/ip will let you fit into 100 mb/s ethernet. but wait, it's been done. That's exactly what HDBaseT alliance does. They claim 10 gb/s uncompressed performance, but the physical medium is 100 mb/s ethernet... how do they do that? They claim to scale the meduim to 1 gb/s. 10 isn't even mentioned (likely because it isn't useful or necessary and would drive up the cost)
100 mbit is good enough, you don't need no stinking gold-plated 10G certified Monster-branded crap.
a 5$ patch you make yourself, or buy from a local computer or hardware store (a mom& pop, not a big-box) will do fine.
actually the first bit was correct, the poem transcription is wrong... actual poem:
Whose woods are these I think I know...
So it's even worse than you imagine, people can remember the phrase inaccurately.
It's just an economic argument. Whether you view investment in domains as boon or bane, opening up TLD's will make far more 'land' available... given supply and demand, this should make the price go down... and also make it necessary to own more "land" to make the business profitable.
You say that your business model is viable. The point is just that opening up TLD's reduces that business's viability.
The artificial scarcity created by limiting.TLD's to a select few makes it feasible for domain squatters to squat on typos. The TLD's were completely open (any possible), then the economics of typo-squatting start to suck because they have to pay for far more domains to hit a lucrative one.
If, say, IBM consolidated to a single TLD (.ibm), they would keep some other domains for legacy purposes, but more than likely, it's harder to typo-squat a TLD than a sub-domain because there are
fewer letters to mess with. Spelling IBM imperceptibly wrong is harder than messing with ibm.com.
legitimate large organizations can better protect their name with access to TLD's than when they are stuck in the.com soup to nuts category.
Concentrating on protecting a smaller number of more unique domains is a net win for legitimate people.
Another thing that might help is some sort of reputation mechanism for DNS. Today we have binary blacklists, it would be helpful to have grey lists indicating domain squatters etc... to give people the option to filter them out. Obviously, these lists would have to be maintained by third parties, such as today's blacklist operators, to be reliable. Augmenting DNS would be cleaner than the current mechanism, and would allow to differentiate amoung DNS providers.
I think TLD's should be used to organize fraud to make it easier to manage. We just ask all tricksters use the.con TLD.
similarly, people with intent to bomb should be steered towards a.terror TLD. or perhaps a more generic.violence.
Also, TLD's should be translated, so that in French the TLD is http://alqaeda.terreur/ if your language setting is FR.
regardless of wether the URL is.con,.fraude,.terreur,.terror,.violence,.violencia, etc... We would then have a simple means of implementing http://www.faqs.org/rfcs/rfc3514.html RFC 3514, just set the evil bit on all traffic coming from these domains.
"Continuous 24 hours-operation can be achieved by placing tight water-filled tubes orbags under the roof. The water heats up during day-time andreleases its heat at night. "
http://www.sbp.de/de/fla/contact/download/The_Solar_Updraft.pdf
This was tested in La Mancha in the 80's.
on the mileage...
You said petrol, so I assume you're in the UK.
you are aware that imperial gallons are 4.22 litres while American gallons are only 3.88 ?
(from a Canadian... the only ones who know about both...)
That's the case right now. Random corporate entity with a strong interest in results will just ask itself:
Who picks the engineers to be on the working groups?
Who pays the engineers? Who pays those who pick the engineers?
OOXML anyone?
funny you should mention the app. We replaced 750 klines of application in TAL, with 20,000 lines of python, which is precisely a 98% reduction in code size. Yes, it took a couple of years, this is mission critical stuff. tackled one functionality facet at a time.
We had a 700 kline app written in some Tandem specific application
language. the smallest server we could get from HP was 400 K$.
we re-wrote the app in python to use pairs of servers replicating via DRDB over ethernet and a load balancer in front. DRBD is slow,
but with the new app I could just add pairs of nodes.
We already had such a configuration for another application, and we
combined the two, so the hardware cost was just adding two nodes in this cluster, at about 4 K$ per server node.
400 K$ -> 8 k$.
I think it would take a heck of a lot of hardware to compensate for the pricing of that gear.
A lot of folks misunderstanding does not make them correct.
X is the name of the thing. It is a window system. There should be no 's' at the end. "Window" is not part of the name. Look at any copyright notice, or any related site, like: http://www.xfree86.org/
"Home of the X Window System." The display server is an X Server, not an X-Windows Server. the clients are X clients, not X-windows clients. Look at the wikipedia article, it refers to the X protocol, X terminals, ad nauseam...
So yeah, there should darn well better be a [sic].
Yes, exactly. X11 ran reasonably complicated applications 20 years ago on hardware that we throw out as woefully inadequate (or quaintly archaic) today, and did so with entirely acceptable speed. X11 isn't the problem -- hardware is what, two orders of magnitude faster now?
20 years of Moore's law is 2**10 == 1024 so... three orders of magnitude.
You sir, are clue impaired. Between
90 and 97% of email on the internet is spam. We are using roaring penguin CANIT pro (basically commercially supported MIMEdefang) as a corporate spam filter. We have the ability to turn off all filtering for clients that ask for it. I love that. There used to be lots people who claimed such things, and we are able to oblige them. They always change their mind within a single day. They clamour to get the filtering back as soon as possible, and then thank me profusely for the effective filters we have in place. Could not ask for better marketing... And I don't have to argue with the naive users, excuse me... upper management. win-win.
The costs is that I have to accept the connections and the email, and drop it using filtering rules later. Since blacklists do around 60-80% of the blocking, that is a huge cpu load, but given the benefit of only a handful of executives experience of actual internet traffic, even if it doubles the number of servers, it easily pays for itself.
People over 65 in the 20's were SOL. If you got rid of social security and Medicare and medicaid, you could cut your taxes in half. Include the military, and the above covers 80% of the federal budget. So cut the military 90%, and eliminate the other two, and you'll be back to the good old 20's. Good luck getting elected on that platform...
One of the biggest reasons not to live in the US is healthcare. My government monopoly (in Canada) lets me go to any doctor I want. and we pay less than half, per capita, what Americans spend on health care, which means that those with coverage are likely spending three times as much or more. Yet Canadians live longer... maybe it's just that the poor Americans are dying like flies, bringing the average down... you're going to try to spin that into a positive, aren't you?
Americans have health care that is great when you are not sick, which like a roof that doesn't leak when it isn't raining. I think many of those who have health insurance will get surprised if they ever actually have the audacity to fall seriously ill.
Why are you spending so much money, for so little care? Maybe because there are no competitive incentives to control costs. The higher the premiums are, the higher the justifiable management costs. The only reason to reduce costs in the US system is to improve the profit margin, but one can do the same by just raising the price, which has the added benefit of helping the HMO too, so it's pretty moot.
The market doesn't work, if only healthy people can shop for health insurance. You cannot test drive it, you're basically buying on reputation, and how they take care of your minor issues and checkups while you are healthy.
Once you are sick, you can be very sure that you cannot change providers. You can be very sure that they are looking for reasons to unload you, that you are at higher risk of being fired because the health insurance provider squeezes your employer. Insurance companies naturally seek to cover the least risk for the highest price.
There is no free market when you are sick in a private system, Your insurance provider, at the time you fall ill is essentially the only game in town. Can you say Monopoly? knew you could! When there is public health care (ie. a public system to pay private doctors to provide services (which is how it works in Canada)) The patient can change doctors at any point, even after they are sick, get a second or third opinion any time they want, and do not have to worry about losing their job.
In public health insurance, risk is just removed as an economic factor, and is distributed over the entire population. People pay a flat rate for health care. The doctors are still private, The hospitals can be as well. The fees are standardized, so the doctors compete on cost reduction (to maximize their profits.) Since the fees are negotiated by the government in bulk, everyone benefits from wholesale-style pricing. That's a big reason why drugs in Canada cost a fraction of what they do in the US.
But you know, just keep on keeping on, because it's better for us. Canadian labour costs are lower (for example in the auto industry) because the health insurance premiums for employees are 1/4 what they are in the US, including for the large number or retirees. That's one big reason why plants are closing in the US and relocating... to Canada... http://www.cbc.ca/canada/story/2009/08/28/toyota-corolla-cambridge592.html
Your claim would appear to be that, as soon as any change in encoding is introduced, it ceases to be comparable. VOIP is clearly impossible then, when folks deigned to replace 56 khz analog lines with many voice coding regimes, many of them negotiated by both parties using protocols such as SIP, and they clearly use a lot less than 56 Khz, must be completely unusable.
your numbers for 1080p: 32bpp x 1920 x 1080 pixels x 24 fps = 1.5 Gbps. raw. You consider this an insurmountable problem.
For video, you claim that any compression is not possible. In contrast to past experience with voice, in spite of the well known source data (both OTA and disk formats) using MPEG compression to routinely achieve between 15x and 30x reduction in volume, ( http://en.wikipedia.org/wiki/Video_compression)
We started this discussion with you claiming that it adds extra hardware. I said, from the beginning, that most sources of high res. data are already mpeg: OTA HD transmission, DVD disk content being examples. Those cases require no extra hardware at all. I grant you both that menu generation is a complication, and that Game Consoles would indeed require a little beefier chip somewhere. You say that this is expensive. I say it would be cheap. I think we have said this often enough to just agree to disagree.
We started this discussion with you claiming an 8-port HDMI switch is more effective and cheaper than ethernet. This last post said that HDMI isn't anywhere close to replacing ethernet. You might want to make up your mind about what you are trying to claim, or you might sound asinine.
Since the outset on this side, the point being proposed is the opposite of what you describe: whenever the opportunity arises, if one can replace an application-specific connection with a more robust general purpose one that is a win. This is similar to iSCSI or FCoE functionally replacing SCSI and FC cabling for low end applications. Instead of having your own application specific, baked in hardware protocol and connectors, leave the hardware to something existing and robust, and build the application specific stuff as an application layer over ethernet. That's iSCSI in a nutshell.
I think we're at the point where HDMIoE would be a smart thing to do. You think it's expensive. I think we're done here.
Slashdot Mirror
http://www.tuxradar.com/content/control-your-bandwidth-trickle
It's done! by google, no less...
Have a little config file consulted by the shim library:
~/.bwnice.conf: 199.237.54.1:80 5 MB/s /tmp/videocache.bin 5 MB/s
etc...
I dunno what your use case is, you could use this for stuff you start as a user, or added to the system config, and have it apply to all users. To do system-wide stuff, tc is already plenty good enough, thought it only applies to network stuff. Dunno of anything for limiting b/w to local disk.
anyway, you don't need any kernel anything for this feature. basic rule is if it can be done in user space, it very often ought to be done there.
http://www.ibm.com/developerworks/linux/library/l-glibc.html
http://www.linuxjournal.com/article/7795
The problem is that the "IT Architecture" used by most organisations is "something that works on what we have now" without the slightest appreciation of how ephemeral "now" is. Most organizations don't give a rats behind for open standards, and the result is the common rats nest, which is paradise for contractors and vendors. It's lock-in through entanglement.
Those people who refused are very rational, and very trapped. Their problem is the result of rampant ignorance and very poor decision making, but it is typical of many large organizations.
But on the other hand...All video signals can/are compressed. The over the air standard is MPEG2 compressed 1080i which fits in 6 MHZ of bandwidth. as in about 6 mbits... as in 1/200th of 10 Gbit/s. All TV's have MPEG decoders because, well, that's what they need to be TV's. Just standardizing compression over tcp/ip will let you fit into 100 mb/s ethernet. but wait, it's been done. That's exactly what HDBaseT alliance does. They claim 10 gb/s uncompressed performance, but the physical medium is 100 mb/s ethernet... how do they do that? They claim to scale the meduim to 1 gb/s. 10 isn't even mentioned (likely because it isn't useful or necessary and would drive up the cost)
100 mbit is good enough, you don't need no stinking gold-plated 10G certified Monster-branded crap. a 5$ patch you make yourself, or buy from a local computer or hardware store (a mom& pop, not a big-box) will do fine.
Dear UserCrisCanter4 http://slashdot.org/comments.pl?sid=1251091&cid=28216951 It only took a year.
actually the first bit was correct, the poem transcription is wrong... actual poem:
Whose woods are these I think I know...
So it's even worse than you imagine, people can remember the phrase inaccurately.
It's just an economic argument. Whether you view investment in domains as boon or bane, opening up TLD's will make far more 'land' available... given supply and demand, this should make the price go down... and also make it necessary to own more "land" to make the business profitable. You say that your business model is viable. The point is just that opening up TLD's reduces that business's viability.
The artificial scarcity created by limiting .TLD's to a select few makes it feasible for domain squatters to squat on typos. The TLD's were completely open (any possible), then the economics of typo-squatting start to suck because they have to pay for far more domains to hit a lucrative one.
If, say, IBM consolidated to a single TLD (.ibm), they would keep some other domains for legacy purposes, but more than likely, it's harder to typo-squat a TLD than a sub-domain because there are
fewer letters to mess with. Spelling IBM imperceptibly wrong is harder than messing with ibm.com.
legitimate large organizations can better protect their name with access to TLD's than when they are stuck in the .com soup to nuts category.
Concentrating on protecting a smaller number of more unique domains is a net win for legitimate people.
Another thing that might help is some sort of reputation mechanism for DNS. Today we have binary blacklists, it would be helpful to have grey lists indicating domain squatters etc... to give people the option to filter them out. Obviously, these lists would have to be maintained by third parties, such as today's blacklist operators, to be reliable. Augmenting DNS would be cleaner than the current mechanism, and would allow to differentiate amoung DNS providers.
I think TLD's should be used to organize fraud to make it easier to manage. We just ask all tricksters use the .con TLD.
similarly, people with intent to bomb should be steered towards a .terror TLD. or perhaps a more generic .violence.
Also, TLD's should be translated, so that in French the TLD is http://alqaeda.terreur/ if your language setting is FR.
regardless of wether the URL is .con, .fraude, .terreur, .terror, .violence, .violencia, etc... We would then have a simple means of implementing http://www.faqs.org/rfcs/rfc3514.html RFC 3514, just set the evil bit on all traffic coming from these domains.
"Continuous 24 hours-operation can be achieved by placing tight water-filled tubes orbags under the roof. The water heats up during day-time andreleases its heat at night. " http://www.sbp.de/de/fla/contact/download/The_Solar_Updraft.pdf This was tested in La Mancha in the 80's.
on the mileage... You said petrol, so I assume you're in the UK. you are aware that imperial gallons are 4.22 litres while American gallons are only 3.88 ? (from a Canadian... the only ones who know about both...)
python says: 6**0.5 = 2.449... So if a planet with 6 times the mass has 2.45 times the diameter at the surface, It's juuussst right!
Especially when the one state is Nevada
That's the case right now. Random corporate entity with a strong interest in results will just ask itself: Who picks the engineers to be on the working groups? Who pays the engineers? Who pays those who pick the engineers? OOXML anyone?
funny you should mention the app. We replaced 750 klines of application in TAL, with 20,000 lines of python, which is precisely a 98% reduction in code size. Yes, it took a couple of years, this is mission critical stuff. tackled one functionality facet at a time.
We had a 700 kline app written in some Tandem specific application language. the smallest server we could get from HP was 400 K$. we re-wrote the app in python to use pairs of servers replicating via DRDB over ethernet and a load balancer in front. DRBD is slow, but with the new app I could just add pairs of nodes. We already had such a configuration for another application, and we combined the two, so the hardware cost was just adding two nodes in this cluster, at about 4 K$ per server node. 400 K$ -> 8 k$. I think it would take a heck of a lot of hardware to compensate for the pricing of that gear.
Which is a European colony, and Finland is part of the EC. Which means it all came from Europe.
everything you need for an all night hack! Thanks America!
in other news, Canadians invented basketball, the minivan, the donut, and perfected bacon. Where would y'all be without us!
X is the name of the thing. It is a window system. There should be no 's' at the end. "Window" is not part of the name. Look at any copyright notice, or any related site, like: http://www.xfree86.org/ "Home of the X Window System." The display server is an X Server, not an X-Windows Server. the clients are X clients, not X-windows clients. Look at the wikipedia article, it refers to the X protocol, X terminals, ad nauseam ...
So yeah, there should darn well better be a [sic].
20 years of Moore's law is 2**10 == 1024 so ... three orders of magnitude.
You sir, are clue impaired. Between 90 and 97% of email on the internet is spam. We are using roaring penguin CANIT pro (basically commercially supported MIMEdefang) as a corporate spam filter. We have the ability to turn off all filtering for clients that ask for it. I love that. There used to be lots people who claimed such things, and we are able to oblige them. They always change their mind within a single day. They clamour to get the filtering back as soon as possible, and then thank me profusely for the effective filters we have in place. Could not ask for better marketing... And I don't have to argue with the naive users, excuse me... upper management. win-win. The costs is that I have to accept the connections and the email, and drop it using filtering rules later. Since blacklists do around 60-80% of the blocking, that is a huge cpu load, but given the benefit of only a handful of executives experience of actual internet traffic, even if it doubles the number of servers, it easily pays for itself.
The reason there is income tax is folks ask the government to do more than it did in the past.
Please consider size of the U.S. Military in 1920.
http://www.answers.com/topic/u-s-army-1900-41
120,000 men. Today... around 3 million people
http://en.wikipedia.org/wiki/United_States_armed_forces
People over 65 in the 20's were SOL. If you got rid of social security and Medicare and medicaid, you could cut your taxes in half. Include the military, and
the above covers 80% of the federal budget. So cut the military 90%, and eliminate the other two, and you'll be back to the good old 20's. Good luck getting elected on that platform...
One of the biggest reasons not to live in the US is healthcare. My government monopoly (in Canada) lets me go to any doctor I want. and we pay less than half, per capita, what Americans spend on health care, which means that those with coverage are likely spending three times as much or more. Yet Canadians live longer... maybe it's just that the poor Americans are dying like flies, bringing the average down... you're going to try to spin that into a positive, aren't you?
Americans have health care that is great when you are not sick, which like a roof that doesn't leak when it isn't raining. I think many of those who have health insurance will get surprised if they ever actually have the audacity to fall seriously ill.
Why are you spending so much money, for so little care? Maybe because there are no competitive incentives to control costs. The higher the premiums are, the higher the justifiable management costs.
The only reason to reduce costs in the US system is to improve the profit margin, but one can do the same by just raising the price, which has the added benefit of helping the HMO too, so it's pretty moot.
The market doesn't work, if only healthy people can shop for health insurance. You cannot test drive it, you're basically buying on reputation, and how they take care of your minor issues and checkups while you are healthy.
Once you are sick, you can be very sure that you cannot change providers. You can be very sure that they are looking for reasons to unload you, that you are at higher risk of being fired because the health insurance provider squeezes your employer. Insurance companies naturally seek to cover the least risk for the highest price.
There is no free market when you are sick in a private system, Your insurance provider, at the time you fall ill is essentially the only game in town. Can you say Monopoly? knew you could! When there is public health care (ie. a public system to pay private doctors to provide services (which is how it works in Canada)) The patient can change doctors at any point, even after they are sick, get a second or third opinion any time they want, and do not have to worry about losing their job.
In public health insurance, risk is just removed as an economic factor, and is distributed over the entire population. People pay a flat rate for health care.
The doctors are still private, The hospitals can be as well. The fees are standardized, so the doctors compete on cost reduction (to maximize their profits.)
Since the fees are negotiated by the government in bulk, everyone benefits from wholesale-style pricing. That's a big reason why drugs in Canada cost a fraction of what they do in the US.
But you know, just keep on keeping on, because it's better for us. Canadian labour costs are lower (for example in the auto industry) because the health insurance premiums for employees are 1/4 what they are in the US, including for the large number or retirees. That's one big reason why plants are closing in the US and relocating... to Canada...
http://www.cbc.ca/canada/story/2009/08/28/toyota-corolla-cambridge592.html
From a human pers
Your claim would appear to be that, as soon as any change in encoding is introduced, it ceases to be comparable. VOIP is clearly impossible then, when folks deigned to replace 56 khz analog lines with many voice coding regimes, many of them negotiated by both parties using protocols such as SIP, and they clearly use a lot less than 56 Khz, must be completely unusable.
your numbers for 1080p:
32bpp x 1920 x 1080 pixels x 24 fps = 1.5 Gbps.
raw. You consider this an insurmountable problem.
For video, you claim that any compression is not possible. In contrast to past experience with voice, in spite of the well known source data (both OTA and disk formats) using MPEG compression to routinely achieve between 15x and 30x reduction in volume, (
http://en.wikipedia.org/wiki/Video_compression)
If it's not HDMIoE, then it's at leas ViOIP.
We started this discussion with you claiming that it adds extra hardware. I said, from the beginning, that most sources of high res. data are already mpeg: OTA HD transmission, DVD disk content being examples. Those cases require no extra hardware at all. I grant you both that menu generation is a complication, and that Game Consoles would indeed require a little beefier chip somewhere. You say that this is expensive. I say it would be cheap. I think we have said this often enough to just agree to disagree.
We started this discussion with you claiming an 8-port HDMI switch is more effective and cheaper than ethernet. This last post said that HDMI isn't anywhere close to replacing ethernet. You might want to make up your mind about what you are trying to claim, or you might sound asinine.
Since the outset on this side, the point being proposed is the opposite of what you describe: whenever the opportunity arises, if one can replace an application-specific connection with a more robust general purpose one that is a win. This is similar to iSCSI or FCoE functionally replacing SCSI and FC cabling for low end applications. Instead of having your own application specific, baked in hardware protocol and connectors, leave the hardware to something existing and robust, and build the application specific stuff as an application layer over ethernet. That's iSCSI in a nutshell.
I think we're at the point where HDMIoE would be a smart thing to do. You think it's expensive. I think we're done here.