>they could have been testing how well their attack would work
Good insight, but why attack the root servers in the first place?
The days when people tried to burn down the Internet just to watch the flames dancing ended a few years ago. It's about profit now. If a crook launches a DDoS on a gambling site the day before the Super Bowl, that crook can extort money. Crooks can also make crooked money from click fraud or spam runs.
Where's the money in taking down the root DNS servers? Why would a crook throw away the black market value of a botnet to do something that wouldn't bring in loot?
Yes. Not only is the wrong question, it doesn't even make sense.
Open source and closed source are methods, security is a result. Security is an attribute of a product, not of a development technique. A closed-source company can assign a hundred reviewers and get more trained eyeballs on their code than most open source projects ever see.
If you want to measure results, there's so much scatter from other causes that any effects of open vs. closed are swamped in the noise. Which would you pick as an example of open source security -- OpenBSD, or sendmail? Which would you pick as an example of closed source security -- VMS, or Internet Explorer?
If you've made all your other security-related decisions and then decide whether to publish source code, one thing to consider is how motivated the attackers are. The crypto community is wedded to published algorithms because they have to face attackers with national budgets who can hire people like Alan Turing. Vertical market software for running the environmental controls in a chicken coop doesn't need, and won't get, worldwide peer review.
There's also the design-to-failure argument. Secrecy is fragile and temporary, and repairs are difficult when it's lost.
>Were the Bush administration not to have gone tax-cut-happy the moment it came to power, NASA might have slightly more resources at its disposal.
The current Aviation Week has an editorial about the budget negotiations. The Administration's Office of Management and Budget is trying to get Congress to reconsider cuts for military bases, energy initiatives, veteran's affairs, Social Security, something called the American Competitiveness Initiative, AIDS expenses, something called the Millenium Challenge Corporation, something called the Economic Support Fund, the DOJ, the General Services Administration, Amtrak, and has raised questions about the budget resolution's language on labor regulations at a "mixed oxide fabrication facility" in South Carolina.
What have they said to Congress about NASA? *crickets chirping*
There are several worthy causes on that list but it speaks volumes that the Administration does not consider NASA a worthy cause.
>and by the time that happens, the criminal is long gone.
Police can and do get warrants over the phone, waking up judges if needed and swearing to the facts that constitute probable cause. The need to get warrants hasn't stopped the police and courts from putting over 2 million USians in prison.
>the belief that some company supplying some cryptographic technology has people in it who are smarter than everybody else in the world?
Easy solution. Let everyone else in the world look at the algorithm. If people smarter than you can't break it or even put hairline cracks into it after several years of trying, then trusting it is a sounder bet than you made the last time you took a job or got married. AES was the output of such a process.
This is exactly why crypto people despise secret and proprietary crypto schemes. Anybody using one of those is betting that they, or the people they hire, are the smartest people in the world. Statistically unlikely.
For high-end passwords I've been steering people toward five- or six-word Diceware passphrases. If physical dice are completely random, then that's 64.5 or 77.3 bits of entropy. An attacker could read them out of swap space, plant a keylogger, or analyze the timing of your keyclicks, but they're outside the reach of clever guessing or feasible brute force.
All beside the point, because that's not what the Register article was about, nor was it what the forensics types were talking about. The big point here is *crypto does not solve problems*. Crypto moves problems around. It turns the problem of protecting data into the problem of protecting keys. You hope it's an easier problem, but if you don't protect keys (keep the nerdstick on a chain around your neck, don't have a recovery key, etc.) then it's the same as not protecting your data. Same issue as with PGP: the easy way into PGP-encrypted files is to brute-force the idiotic passphrase that's almost certainly being used to conceal the private key.
Banks have been aware of insider threats for centuries. They have a battle-tested set of policies and procedures such as separation of duties to control the threats. Banks have been able to stay in business for a long time before ESM became available.
Banks have also gone out of business due to the insider threat people seem afraid to discuss. There's an old saying, "The best way to rob a bank is to own one". Crooked senior management stole one Sagan (billions and billions) of dollars during the 1980s US savings and loan disaster. Sometimes the thefts are even considered legal, as when a CEO walks away from a ruined company with a hundred million in "performance bonuses". How is ESM going to protect against Ken Lay, who did more damage than any random thousand "disgruntled former employees"? (*)
Banking procedures, such as requiring people to take vacations, have the other advantage that they don't risk violating privacy laws. In some countries you may not be allowed to spy on your workers to the extent you can in others.
The court looked at various precedents and noticed that what the police were doing was *not* really like any of them, and so needed separate legal authorization and separate thinking-through.
Germany has stricter privacy laws, more passionately enforced, than the UK/US, but this decision is completely compatible with UK/US law that says the scope of a search has to be explicitly defined and minimal. Spyware on a computer fits neither criterion.
>>"Limited support for multi-threading That was hardly even a relevant argument in 1993 but it meant that Symbian uses 'active objects' instead of threads in almost all applications." >In fact, the cost of a OS context-switch is still high when every bit of battery power matters - battery technology hasn't changed that much since 1993 Say more, please? A thread context switch is simple and quick. Even a context switch for a full heavyweight process doesn't change the display's power draw or (shudder) turn on the RF circuitry. How is cooperative multitasking cheaper than a threaded programming model?
>However, there is a non-carbon material harder than diamond (ultrahard fullerine)
Did you mean "non-diamond"? It's still a carbon based material, C60 polymerized.
Carbon. Is there anything it can't do?
Because people forget experience
on
Why Software is Hard
·
· Score: 3, Informative
Fred Brooks had much the same material in _The Mythical Man-Month_: communication overhead spirals out of control in large groups, project scope creeps out to infinity without a budget, overconfident people try to do too much and fail, it's impossible to know what the customer wants and (in a new area) even what works until you've built something and watched how it fails, only make change to known-good baselines, etc.
This author had to discover Fred Brooks after he'd started a career of big projects. TMM should have been in his school curriculum.
If you're OK with just working with physics as opposed to doing research yourself, there are other places where software is a research tool and the people who write it get to learn about the physics.
The national supercomputer centers sometimes have ultimately cool projects like simulating galaxies colliding and rendering images of the result. That gets you the fun of programming big iron, some really challenging numerical problems, shock wave physics, and the chance to watch science being made from the inside. Once there, you can at least try to impress the right people and move into doing your own research. For sure you'll get a clearer idea whether that's what you want to do.
>Just because ExxonMobil paid someone, does not mean the arguments the scientist made are not valid,
If the arguments are valid, they'll show up without bribes ("publish or perish"). ExxonMobil is a business. If they spend money it's because they need to. If the papers they want would show up without bribes, they wouldn't budget money to produce them. They are budgeting money.
>ad hominem attacks.
1. ExxonMobil only spends money to make things happen that wouldn't happen otherwise. 2. ExxonMobil is spending money to make scientific papers happen with a particular agenda. Therefore the papers wouldn't happen without the money.
1. Only invalid arguments require money from ExxonMobil 2. The papers ExxonMobil wants won't appear without money from ExxonMobil. Therefore the arguments are invald.
Slashdot Mirror
>the internet was originally designed to run as a communications network in the event of a nuclear attack.
g y.html
That idea was floating around but it wasn't what drove the MIT/DARPA work that turned into today's Internet:
http://www.ziplink.net/~lroberts/InternetChronolo
>they could have been testing how well their attack would work
Good insight, but why attack the root servers in the first place?
The days when people tried to burn down the Internet just to watch the flames dancing ended a few years ago. It's about profit now. If a crook launches a DDoS on a gambling site the day before the Super Bowl, that crook can extort money. Crooks can also make crooked money from click fraud or spam runs.
Where's the money in taking down the root DNS servers? Why would a crook throw away the black market value of a botnet to do something that wouldn't bring in loot?
South Korea has great residential broadband. It must be a premium place to recruit zombies.
Yes. Not only is the wrong question, it doesn't even make sense.
Open source and closed source are methods, security is a result. Security is an attribute of a product, not of a development technique. A closed-source company can assign a hundred reviewers and get more trained eyeballs on their code than most open source projects ever see.
If you want to measure results, there's so much scatter from other causes that any effects of open vs. closed are swamped in the noise. Which would you pick as an example of open source security -- OpenBSD, or sendmail? Which would you pick as an example of closed source security -- VMS, or Internet Explorer?
If you've made all your other security-related decisions and then decide whether to publish source code, one thing to consider is how motivated the attackers are. The crypto community is wedded to published algorithms because they have to face attackers with national budgets who can hire people like Alan Turing. Vertical market software for running the environmental controls in a chicken coop doesn't need, and won't get, worldwide peer review.
There's also the design-to-failure argument. Secrecy is fragile and temporary, and repairs are difficult when it's lost.
>anything it would collide with would also be moving at a comparable speed
Comparable speed but not comparable velocity: if something in polar orbit hits something in equatorial orbit, grief will ensue.
>they will bring in the bomb squad and they will treat it very seriously.
In the days of the Unabomber, a university researcher (my future wife) got a package with no return address. She called it in.
The responding officer said "It doesn't seem like a bomb" as she shook it.
>Were the Bush administration not to have gone tax-cut-happy the moment it came to power, NASA might have slightly more resources at its disposal.
The current Aviation Week has an editorial about the budget negotiations. The Administration's Office of Management and Budget is trying to get Congress to reconsider cuts for military bases, energy initiatives, veteran's affairs, Social Security, something called the American Competitiveness Initiative, AIDS expenses, something called the Millenium Challenge Corporation, something called the Economic Support Fund, the DOJ, the General Services Administration, Amtrak, and has raised questions about the budget resolution's language on labor regulations at a "mixed oxide fabrication facility" in South Carolina.
What have they said to Congress about NASA? *crickets chirping*
There are several worthy causes on that list but it speaks volumes that the Administration does not consider NASA a worthy cause.
>You cannot just say whatever you want in a newspaper or in a public forum without people knowing who you are.
"Name witheld by request"
"A source who asked to be anonymous"
Anyone handing out pamphlets on the street corner
The authors of the Federalist Papers.
>and by the time that happens, the criminal is long gone.
Police can and do get warrants over the phone, waking up judges if needed and swearing to the facts that constitute probable cause. The need to get warrants hasn't stopped the police and courts from putting over 2 million USians in prison.
>the belief that some company supplying some cryptographic technology has people in it who are smarter than everybody else in the world?
Easy solution. Let everyone else in the world look at the algorithm. If people smarter than you can't break it or even put hairline cracks into it after several years of trying, then trusting it is a sounder bet than you made the last time you took a job or got married. AES was the output of such a process.
This is exactly why crypto people despise secret and proprietary crypto schemes. Anybody using one of those is betting that they, or the people they hire, are the smartest people in the world. Statistically unlikely.
For high-end passwords I've been steering people toward five- or six-word Diceware passphrases. If physical dice are completely random, then that's 64.5 or 77.3 bits of entropy. An attacker could read them out of swap space, plant a keylogger, or analyze the timing of your keyclicks, but they're outside the reach of clever guessing or feasible brute force.
Bitlocker uses AES-CBC with some tweakage to deter the usual attacks against full-disk encryption with CBC. The Microsoft paper about encryption options for full-disk encryption is really not bad.
All beside the point, because that's not what the Register article was about, nor was it what the forensics types were talking about. The big point here is *crypto does not solve problems*. Crypto moves problems around. It turns the problem of protecting data into the problem of protecting keys. You hope it's an easier problem, but if you don't protect keys (keep the nerdstick on a chain around your neck, don't have a recovery key, etc.) then it's the same as not protecting your data. Same issue as with PGP: the easy way into PGP-encrypted files is to brute-force the idiotic passphrase that's almost certainly being used to conceal the private key.
Banks have been aware of insider threats for centuries. They have a battle-tested set of policies and procedures such as separation of duties to control the threats. Banks have been able to stay in business for a long time before ESM became available.
Banks have also gone out of business due to the insider threat people seem afraid to discuss. There's an old saying, "The best way to rob a bank is to own one". Crooked senior management stole one Sagan (billions and billions) of dollars during the 1980s US savings and loan disaster. Sometimes the thefts are even considered legal, as when a CEO walks away from a ruined company with a hundred million in "performance bonuses". How is ESM going to protect against Ken Lay, who did more damage than any random thousand "disgruntled former employees"? (*)
Banking procedures, such as requiring people to take vacations, have the other advantage that they don't risk violating privacy laws. In some countries you may not be allowed to spy on your workers to the extent you can in others.
(*) Who disgruntled them, anyway?
The court looked at various precedents and noticed that what the police were doing was *not* really like any of them, and so needed separate legal authorization and separate thinking-through.
Germany has stricter privacy laws, more passionately enforced, than the UK/US, but this decision is completely compatible with UK/US law that says the scope of a search has to be explicitly defined and minimal. Spyware on a computer fits neither criterion.
>>"Limited support for multi-threading That was hardly even a relevant argument in 1993 but it meant that Symbian uses 'active objects' instead of threads in almost all applications."
>In fact, the cost of a OS context-switch is still high when every bit of battery power matters - battery technology hasn't changed that much since 1993
Say more, please? A thread context switch is simple and quick. Even a context switch for a full heavyweight process doesn't change the display's power draw or (shudder) turn on the RF circuitry. How is cooperative multitasking cheaper than a threaded programming model?
>However, there is a non-carbon material harder than diamond (ultrahard fullerine)
Did you mean "non-diamond"? It's still a carbon based material, C60 polymerized.
Carbon. Is there anything it can't do?
Fred Brooks had much the same material in _The Mythical Man-Month_: communication overhead spirals out of control in large groups, project scope creeps out to infinity without a budget, overconfident people try to do too much and fail, it's impossible to know what the customer wants and (in a new area) even what works until you've built something and watched how it fails, only make change to known-good baselines, etc.
This author had to discover Fred Brooks after he'd started a career of big projects. TMM should have been in his school curriculum.
If you're OK with just working with physics as opposed to doing research yourself, there are other places where software is a research tool and the people who write it get to learn about the physics.
The national supercomputer centers sometimes have ultimately cool projects like simulating galaxies colliding and rendering images of the result. That gets you the fun of programming big iron, some really challenging numerical problems, shock wave physics, and the chance to watch science being made from the inside. Once there, you can at least try to impress the right people and move into doing your own research. For sure you'll get a clearer idea whether that's what you want to do.
>hiring scientists to document any and all shortcomings for them
If that's what they wanted, they could get it for free. It's called "peer review".
Funny you should compare Big Tobacco with Big Oil's tactics.
>grant money that only comes in if they say the exact opposite.
A lot of the research is funded by the US federal government.
Do you assert that the Bush administration is determined, regardless of facts, to prove that burning fossile fuels is harmful?
>Just because ExxonMobil paid someone, does not mean the arguments the scientist made are not valid,
If the arguments are valid, they'll show up without bribes ("publish or perish"). ExxonMobil is a business. If they spend money it's because they need to. If the papers they want would show up without bribes, they wouldn't budget money to produce them. They are budgeting money.
>ad hominem attacks.
1. ExxonMobil only spends money to make things happen that wouldn't happen otherwise.
2. ExxonMobil is spending money to make scientific papers happen with a particular agenda.
Therefore the papers wouldn't happen without the money.
1. Only invalid arguments require money from ExxonMobil
2. The papers ExxonMobil wants won't appear without money from ExxonMobil.
Therefore the arguments are invald.
http://www.panexa.com/
"Panexa: ask your doctor for a reason to take it"
>no one was rounded up
Peter Berdovsky has been arrested and charged with a felony.
This, while unidentified trucks are allowed into "secure" areas of Logan Airport. People who can't assess threats can't protect against them.