Both of these guys are very very deserving of an award like this. Linux wouldn't be where it is without Don Becker, and without Guido, there would be no Python! (I'm a Python fan:)
Why are you saying 'we need to start developing cryptography algorithms which aren't factoring based, and we need to start NOW'...as if we don't already have plenty such algorithms. Look into ElGamal or any other discrete log cryptosystem. ElGamal has the bonus have having an already-expired patent...
Hehe...just because you don't know about them doesn't mean there aren't a load of them:)
You have ElGamal, based on discrete logs...(and the discrete log problem generalizes easily to allow for many many many variants on the same principle). DSA is based on discrete logs, too.
Most EC cryptosystems are based on an analog of the the discrete log problem on a finite field...generally a more intractable problem than the general discrete log problem (except for a few degenerate cases, in which it reduces to the problem of discrete log in a finite field).
McEliece, based on algebraeic coding theory...
I think there is one cryptosystem based on the knapsack problem which is yet unbroken (most knapsacks are shown to be weak)
? What you're saying makes no sense...you write as if the government only uses encryption based around 1 principle. It's not like the government says "ok, we're going with discrete logs" and standardizes on that....they pick standards, and the standards use whatever method they happen to use. There is definitely (and has always been) cryptosystems in use within the government that are based on a variety of principles. So, like I said, this makes the Sneakers idea way off and your comment about 'what the governmetn was using at that time' bogus too..
I'm not sure how the weakness of one particular algorithm (RSA) has implies vulnerability in 'modern cryptographic techniques'.
Sure, this shows a weakness in RSA (but if you are using 512-bit keys right now, you are living dangerously as far as what's recommended)
Shamir's discovery doesn't imply any weaknesses in the entire class of public key cryptosystems based around discrete logs. (or other public key cryptosystems not based around prime factorization, such as some knapsack types and McEliece, based on algebraic coding)
The one time pad is impractical for all but a very narrow group of scenarios.
Dude, regardless of advances being made...quantum crypto is *NOWHERE CLOSE* to being practical. Factoring single digit numbers won't get you very far. If you equate 'about to burst onto the scene' with 'several decades', then _maybe_ I might be able to not laugh out loud:)
Uh, do you realize how slow 4096-bit key generation is? Sure, computers are always getting faster...but you have to realize that the state of the art machine is not the 'baseline' platform. You also have to realize that these things need to work with cpu-poor devices, like smart cards. EC cryptosystems help this out to a degree, but they are no panacea either.
Sensible key sizes are a much better solution. Using the max key size 'just because it's there' is just wasteful (and slow as hell).
ECC isn't necessarily tied to any given cryptosystem (there are several cryptosystems which use ECC). It's just that the math operations are done over elliptic curves instead of some other finite field...
(and I'm not sure that the statement about no patents is necessarily true)
But yes, ECC _is_ cool...but needs more research. Lots of people would be a little wary of implementing it, since EC cryptosystems haven't gotten as much scrutiny as of yet.
Big difference here...Shamir's new thing just reduces (by a lot, albeit) the computation required to factor RSA. Assuming this just reduces factoring difficulty, a bunch of other public key cryptosystems won't be affected (like ElGamal, which deals with the difficulty of the discrete log problem instead of prime factorization). The Sneakers story revolved around a little box which (in a matter of seconds) cracked _any_ data scrambled with _any_ algorithm.
CJK, frantically trying to finish up his crypto semester project (due tomorrow), which deals with discrete log public key crypto:)
I don't know what's up with Slashdot readers crying 'Beowulf' whenever they see something that needs CPU juice...
***Beowulf clusters are only good for _certain_ types of parallel computing problems*** (coarsely grained problems are what it's for...throw something too finely grained at it, and it kills itself with internode communication latency) No more talk of a Beowulf Quake:)
CJK
Interesting math you have going there...
on
wcarchive Upgraded
·
· Score: 2
Maybe I'm just totally missing something there, but your numbers seem to make no sense whatsoever (especially since your required bandwidth figure was twice as big at 2500 users vs. 5000). It's like this:
- pretend everyone was connected at 28.8 kbps - this is 3.6 K (kilobytes, not bits) per second - if there are 5000 people @ 3.6 K/second:
This is also operating under the fallacy that everyone is on at 28.8 kbps...there will still be some 14.4's...but also lots of 56 kbps...and then lots of people on faster (perhaps from work) links. Heck, I have 1.54 Mbps (T1) in my apartment....just 12 people like me and you're already sucking up more bandwidth than 5000 people @ 28.8...
Ah..._this_ is what I've been hoping to see more of in the comments here (but it's sorely lacking)..someone who has a backbone AND a clue!
The predators prey on those they perceive to be the weakest...if they don't sense weakness, they'll move on. You don't have to like these 'jocks' (? let's not give the impression that if you're good at sports than you're automatically an asshole)...just stick with the golden rule and you'll be fine with the _overwhelming_ majority of them.
Finally...some sense in this pool of whining self pity! Sure, middle school/high school kids can be cruel (and often are), but you guys need to stop whining about it...just because your skin isn't thick enough.
This whole series of articles is kind of annoying...it just baits all of the 'the system sucks, screw The Man, I'm not a cookie cutter kid' comments thrown out by those lacking either the social or mental skills to survive in the jungle...
Sure, the 'the system' isn't perfect...but all this talk of anarchy and other such craziness...get a grip!
All of this talk about the maligned, brilliant outcasts who just 'see things differently' is such bull.
Being a geek in no way means you are necessarily an outcast (and vice versa).
All of these people whining is really just 'I am socially inept and am picked on because of it. I tend only to relate with inanimate objects.' Lots of these people *are* socially inept in high school..and many times their problem subsides as they beef up their social interaction skills as they grow older. Sure, there will always be a shafted few...but the general case is not to be shafted 'just because'. Have some backbone! Stand up! All of this talk of being pushed around and made fun of...is your skin really that thin? Are you that easy to get to? Everyone experiences this to a degree...it's part of growing up and learning to deal with adverse situations (social ones, in this case)...
Finally some level-headed talk...this the-man-is-trying-to-cramp-our-great-toking-it's-r eally-not-bad-for-you bs was getting pretty deep...
Biology...look into it. It blows my mind how many times I've had conversations with people who honestly believe that none of these drugs do anything adverse to your body. Whew...ignorance...a scary thing.
? "How many RC5 keys?" This is one of the dumber things to hold up as a viable benchmark. At least try comparing on specint/specfp values...how many people are buying there machines to crack RC5?? Don't get me wrong, the G3 is nice, but it's doesn't have enough of a speed edge to be compelling in any sense.
I don't know if that's a fair statement... Sure, the architecture of the system may be old school, but that's not why people set up Beowulf clusters...they buy them for the untouchable price/performance for coarsely-grained problems. End of story. Don't worry, though...Linux development won't stand still...we'll see changes in the future to allow for more flexible architectures.
Whatever. These are old school Alphas...who would build a new Alpha 450 now?? Besides...not to knock Linux clusters (our ACM chapter just brought one online), but this is kind of a bad comparison...as this kind of stuff doesn't show the HUGE difference in internodal bandwidth between these two systems. If you get something that needs a lot of talking between nodes going, the Cray would pretty much rape the cluster like no tomorrow...the latency on switched fast Ethernet (even Gbit Ethernet) just can't compare to these whack (and horrendously expensive) supercomputer interconnection systems.
Slashdot Mirror
? The R10000 is pretty old news...the R12K has been out for a while. Neither particularly kick ass.
it's in Python!
CJK
Both of these guys are very very deserving of an award like this. Linux wouldn't be where it is without Don Becker, and without Guido, there would be no Python! (I'm a Python fan :)
CJK
I fear any cryptographer who decides to implement an algorithm based on how cool its name sounds.
This is kind of dumb...it's easy to make up additional or new names as part of your marketing which are catchy/easy to remember.
CJK
Why are you saying 'we need to start developing cryptography algorithms which aren't factoring based, and we need to start NOW'...as if we don't already have plenty such algorithms. Look into ElGamal or any other discrete log cryptosystem. ElGamal has the bonus have having an already-expired patent...
CJK
Hehe...just because you don't know about them doesn't mean there aren't a load of them :)
You have ElGamal, based on discrete logs...(and the discrete log problem generalizes easily to allow for many many many variants on the same principle). DSA is based on discrete logs, too.
Most EC cryptosystems are based on an analog of the the discrete log problem on a finite field...generally a more intractable problem than the general discrete log problem (except for a few degenerate cases, in which it reduces to the problem of discrete log in a finite field).
McEliece, based on algebraeic coding theory...
I think there is one cryptosystem based on the knapsack problem which is yet unbroken (most knapsacks are shown to be weak)
CJK
? What you're saying makes no sense...you write as if the government only uses encryption based around 1 principle. It's not like the government says "ok, we're going with discrete logs" and standardizes on that....they pick standards, and the standards use whatever method they happen to use. There is definitely (and has always been) cryptosystems in use within the government that are based on a variety of principles. So, like I said, this makes the Sneakers idea way off and your comment about 'what the governmetn was using at that time' bogus too..
CJK
There are many public key cryptosystems which do not get their strength from the difficulty of factoring prime numbers.
CJK
I'm not sure how the weakness of one particular algorithm (RSA) has implies vulnerability in 'modern cryptographic techniques'.
Sure, this shows a weakness in RSA (but if you are using 512-bit keys right now, you are living dangerously as far as what's recommended)
Shamir's discovery doesn't imply any weaknesses in the entire class of public key cryptosystems based around discrete logs. (or other public key cryptosystems not based around prime factorization, such as some knapsack types and McEliece, based on algebraic coding)
The one time pad is impractical for all but a very narrow group of scenarios.
CJK
Dude, regardless of advances being made...quantum crypto is *NOWHERE CLOSE* to being practical. Factoring single digit numbers won't get you very far. If you equate 'about to burst onto the scene' with 'several decades', then _maybe_ I might be able to not laugh out loud :)
CJK
Uh, do you realize how slow 4096-bit key generation is? Sure, computers are always getting faster...but you have to realize that the state of the art machine is not the 'baseline' platform. You also have to realize that these things need to work with cpu-poor devices, like smart cards. EC cryptosystems help this out to a degree, but they are no panacea either.
Sensible key sizes are a much better solution. Using the max key size 'just because it's there' is just wasteful (and slow as hell).
CJK
ECC isn't necessarily tied to any given cryptosystem (there are several cryptosystems which use ECC). It's just that the math operations are done over elliptic curves instead of some other finite field...
(and I'm not sure that the statement about no patents is necessarily true)
But yes, ECC _is_ cool...but needs more research. Lots of people would be a little wary of implementing it, since EC cryptosystems haven't gotten as much scrutiny as of yet.
CJK
Big difference here...Shamir's new thing just reduces (by a lot, albeit) the computation required to factor RSA. Assuming this just reduces factoring difficulty, a bunch of other public key cryptosystems won't be affected (like ElGamal, which deals with the difficulty of the discrete log problem instead of prime factorization). The Sneakers story revolved around a little box which (in a matter of seconds) cracked _any_ data scrambled with _any_ algorithm.
:)
CJK, frantically trying to finish up his crypto semester project (due tomorrow), which deals with discrete log public key crypto
I don't know what's up with Slashdot readers crying 'Beowulf' whenever they see something that needs CPU juice...
:)
***Beowulf clusters are only good for _certain_ types of parallel computing problems*** (coarsely grained problems are what it's for...throw something too finely grained at it, and it kills itself with internode communication latency) No more talk of a Beowulf Quake
CJK
Maybe I'm just totally missing something there, but your numbers seem to make no sense whatsoever (especially since your required bandwidth figure was twice as big at 2500 users vs. 5000). It's like this:
- pretend everyone was connected at 28.8 kbps
- this is 3.6 K (kilobytes, not bits) per second
- if there are 5000 people @ 3.6 K/second:
5000 * 3600 bytes/sec = 18,000,000 bytes/sec = 17.17 megabytes/sec
This is also operating under the fallacy that everyone is on at 28.8 kbps...there will still be some 14.4's...but also lots of 56 kbps...and then lots of people on faster (perhaps from work) links. Heck, I have 1.54 Mbps (T1) in my apartment....just 12 people like me and you're already sucking up more bandwidth than 5000 people @ 28.8...
CJK
Ah..._this_ is what I've been hoping to see more of in the comments here (but it's sorely lacking)..someone who has a backbone AND a clue!
The predators prey on those they perceive to be the weakest...if they don't sense weakness, they'll move on. You don't have to like these 'jocks' (? let's not give the impression that if you're good at sports than you're automatically an asshole)...just stick with the golden rule and you'll be fine with the _overwhelming_ majority of them.
CJK
Finally...some sense in this pool of whining self pity! Sure, middle school/high school kids can be cruel (and often are), but you guys need to stop whining about it...just because your skin isn't thick enough.
This whole series of articles is kind of annoying...it just baits all of the 'the system sucks, screw The Man, I'm not a cookie cutter kid' comments thrown out by those lacking either the social or mental skills to survive in the jungle...
Sure, the 'the system' isn't perfect...but all this talk of anarchy and other such craziness...get a grip!
CJK
Agreed...
All of this talk about the maligned, brilliant outcasts who just 'see things differently' is such bull.
Being a geek in no way means you are necessarily an outcast (and vice versa).
All of these people whining is really just 'I am socially inept and am picked on because of it. I tend only to relate with inanimate objects.' Lots of these people *are* socially inept in high school..and many times their problem subsides as they beef up their social interaction skills as they grow older. Sure, there will always be a shafted few...but the general case is not to be shafted 'just because'. Have some backbone! Stand up! All of this talk of being pushed around and made fun of...is your skin really that thin? Are you that easy to get to? Everyone experiences this to a degree...it's part of growing up and learning to deal with adverse situations (social ones, in this case)...
CJK
Finally some level-headed talk...this the-man-is-trying-to-cramp-our-great-toking-it's-r eally-not-bad-for-you bs was getting pretty deep...
Biology...look into it. It blows my mind how many times I've had conversations with people who honestly believe that none of these drugs do anything adverse to your body. Whew...ignorance...a scary thing.
Agreed.
? "How many RC5 keys?" This is one of the dumber things to hold up as a viable benchmark. At least try comparing on specint/specfp values...how many people are buying there machines to crack RC5?? Don't get me wrong, the G3 is nice, but it's doesn't have enough of a speed edge to be compelling in any sense.
CJK
I don't know if that's a fair statement...
Sure, the architecture of the system may be old school, but that's not why people set up Beowulf clusters...they buy them for the untouchable price/performance for coarsely-grained problems. End of story.
Don't worry, though...Linux development won't stand still...we'll see changes in the future to allow for more flexible architectures.
CJK
Whatever. These are old school Alphas...who would build a new Alpha 450 now?? Besides...not to knock Linux clusters (our ACM chapter just brought one online), but this is kind of a bad comparison...as this kind of stuff doesn't show the HUGE difference in internodal bandwidth between these two systems. If you get something that needs a lot of talking between nodes going, the Cray would pretty much rape the cluster like no tomorrow...the latency on switched fast Ethernet (even Gbit Ethernet) just can't compare to these whack (and horrendously expensive) supercomputer interconnection systems.
CJK