PS3, Linux Used to Catch Child Pornographers

Page 3 - Seeking answers? Join the Tom's Guide community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.

bjforte

Distinguished
Apr 19, 2008
1
0
18,510
Most encryption is designed to be quick and transparent. Its not very difficult to use a massive random key to protect your data, if the key its truely random then once that key is deleted the data is lost forever.
 
G

Guest

Guest
[quote\]I also have a hard time believing 60 PS3s are a better choice than a few fast, but regular computers with a bunch of cores [/quote]

Better believe it. The cell is a RISC processor, which means its extremely good at repetitive tasks, to put it in layman's terms. x86 processors, although fast, are not even close to the cell in terms of sheer number-crunching prowess. The cell behaves more like a GPU in this area. Its not good for general purpose stuff like the x86 processors, but miles ahead in HPC area.

Compared to other RISC processors in the market, like the SPARC, HP PA RISC or the IBM PowerPC, the cell bests them in bandwidth, cores and cost-effectiveness. This is the reason we seee all thee applications for the cell based PS3. As an example, my tyan based server with 2 X quad core 2.4 GHz Shanghai processors takes longer to perform a computation than my PS3. However, when I run the GPU based Folding@Home on my PC, its faster than my PS3. Hope that puts things in perspective for you.
 
G

Guest

Guest
If what they say about CUDA technology is so good, then why not use it?
 

jellico

Distinguished
Apr 17, 2009
412
0
18,930
[citation][nom]diss3nt_iz_patri0tic[/nom]By "child pornographers", I'm sure they mean:political dissidentstax evadersthose who plan on refusing the H1N1 vaccineanybody who disagrees with the direction our country is going inotherwise, they could better spend their time, money and resources spying on the clergy looking for pedo-folks[/citation]
I suspect there's an element of truth to that statement. In any case, human beings have always been, and will always be, the weakest link in computer security. They will install military-grade encryption software, and then choose a simple password that can be brute forced in no time.

The technology does not exist to break most of the modern encryption algorithms. And once you start getting into 128-bits and beyond, you can forget a straight brute-force attack, even with a million networked PS3s. However, what law enforcement agencies have done is focus on the weak link. When they execute a search warrant (or an illegal search, as the case may be), they not only take all of the computers and media, but get as much information about the target of the search as well. They will build a profile that can be fed into a special hacking program that uses a dictionary attack in conjunction with the target profile (all passwords he or she has been known to use, birthdates of all friends and family, addresses, phone numbers, SSNs, etc.). With enough time and processing power, they can crack most people's passwords. That's why it's important to use pass phrases, not password. And the pass phrase should be something you've never written down, never told anyone, and means something ONLY to you. It should be mixed case, spelled phoenetically, lots of special character and number substitutions, etc. If you do that, your password will be secure and damn near unbreakable. Then you only have to worry about side channel attacks (cold boot attack, stoned attack, evil maid attack, etc.)
 

JonathanDeane

Distinguished
Mar 28, 2006
192
0
18,630
I agree with Jellico pass phrases are the way to go these days. Even a modern home PC can brute force passwords all by its lonesome. I use pass phrases myself although I do not go into all the jazz Jellico says to do (of course I have nothing on my machine that would put me in jail I worry more about people getting into my email and using it for spam lol) So a nice pass phrase copied out of a random book works for me and keeps out the riff raff.

Of course as PC's get faster and faster this may change and require pass paragraphs or pass pages.... Then we will be forced to have an entire library on our PC so we can just copy and paste our passwords... (That makes me feel sooo insecure on so many levels lol)

Great example is that even my machine can brute force a Rar password in a couple of days... I wonder how long it will be before encryption can destroy the file if the password is guessed too many times ? (if not now and I just am not aware of it) the file could be copied over and over to allow more guesses but that would slow things down to a crawl even on a RAM disk.
 

Shadow703793

Distinguished
Feb 3, 2007
696
0
18,940
[citation][nom]one-shot[/nom]According to, "http://www.blachford.info/computer/Cell/Cell1_v2.html"SPEs can also be chained, that is they can be set up to process data in a stream using multiple SPEs in parallel. In this mode a Cell may approach its theoretical maximum processing speed of 256 GigaFlops.1-1.8TeraFLOPS sounds a little high for a PS3, no disrespect.[/citation]
Thanks for the correction. So that means, a GPU (like the ATI 5850) has more FLOPs than a CPU and Cell?
 

SlyNine

Distinguished
Feb 1, 2009
10
0
18,560
citation][nom]catalystgod[/nom][quote\]I also have a hard time believing 60 PS3s are a better choice than a few fast, but regular computers with a bunch of cores [/quote]Better believe it. The cell is a RISC processor, which means its extremely good at repetitive tasks, to put it in layman's terms. x86 processors, although fast, are not even close to the cell in terms of sheer number-crunching prowess. The cell behaves more like a GPU in this area. Its not good for general purpose stuff like the x86 processors, but miles ahead in HPC area.Compared to other RISC processors in the market, like the SPARC, HP PA RISC or the IBM PowerPC, the cell bests them in bandwidth, cores and cost-effectiveness. This is the reason we seee all thee applications for the cell based PS3. As an example, my tyan based server with 2 X quad core 2.4 GHz Shanghai processors takes longer to perform a computation than my PS3. However, when I run the GPU based Folding@Home on my PC, its faster than my PS3. Hope that puts things in perspective for you.[/citation]


You also have to understand what kind of WUs each one is crunching, X86 is crunching more complex WUs, while PS3 is crunching more complex ( for now) units then GPUs. Speed at the cost of flexibility.

The PS3 is also highly limited to the platform, With its limited bandwidth and ram it will never come close to its potential.

[citation][nom]JonathanDeane[/nom]. Even a modern home PC can brute force passwords all by its lonesome.[/citation]

The estimated power your computer would consume over some unbelievable period of time would probably prevent it from ever cracking AES 128, nevermind 256. That and we are talking around millions of times the age of the universe if that age is around 13 billion years.

So yea, you could program something to start a brute force attack. But then again you could use have an ante try and move a planet. Neither is going to have success assuming the passphrase isn't the weak link.
 

one-shot

Distinguished
Jan 13, 2006
128
0
18,640
[citation][nom]Shadow703793[/nom]Thanks for the correction. So that means, a GPU (like the ATI 5850) has more FLOPs than a CPU and Cell?[/citation]

All I did was provide a source with information that stated differently from what you claim. You're welcome to post other sources if you find one that's more reliable. GPUs have been known to produce more FLOPS than CPUs. My old 4850 was rated at 1TFLOP. My overclocked i7 isn't quite there.
 

SlyNine

Distinguished
Feb 1, 2009
10
0
18,560
[citation][nom]Shadow703793[/nom]The Cell chip found on the PS3 can easily do 1 (or was it like 1.8?) TFLOPS. On the other hand, a high end Nehalem can do only about 80-120 GFLOPS. The ATI R800s can do ~2.8-3.1 TFLOPS.[/citation]

A 1900XT Would pound the PS3's Cell in raw Gflops. But the Cell is far more useful. Gflops is simply a gross calculation of the maximum calculations per cycle, and operating speed. Now once you throw in real world work the PS3 actually does closer to MAYBE 5-10 Gflops with typical types of programes.

Take Folding@home, 17533 Nvidia GPU's are producing 2086 Tflops., while it takes 34718 PS3's to produce 979 Tflops. Nvidia's GPU's are producing around 110 Gflops, and PS3 is producing around 28 Gflops. Proof that it is certainly NOT true that the PS3 can easily produce 1.8 Tflops. This is for a program that can get really high flops out of the PS3 and GPUs. Also keep in mind we are not talking about high end GPUs, Any GPU that can run Cuda can run Folding@home.

Folding@home is not a typical type of program, it will get much higher Gflops then most.
 

bfstev

Distinguished
Aug 4, 2008
90
0
18,580
I wonder if they realize that 60 ps3s is more expensive than 1 $8000 supercomputer. This just reaks of fail.
 

JonathanDeane

Distinguished
Mar 28, 2006
192
0
18,630
[citation][nom]SlyNine[/nom]citation][nom]catalystgod[/nom][quote\]I also have a hard time believing 60 PS3s are a better choice than a few fast, but regular computers with a bunch of cores [/quote]Better believe it. The cell is a RISC processor, which means its extremely good at repetitive tasks, to put it in layman's terms. x86 processors, although fast, are not even close to the cell in terms of sheer number-crunching prowess. The cell behaves more like a GPU in this area. Its not good for general purpose stuff like the x86 processors, but miles ahead in HPC area.Compared to other RISC processors in the market, like the SPARC, HP PA RISC or the IBM PowerPC, the cell bests them in bandwidth, cores and cost-effectiveness. This is the reason we seee all thee applications for the cell based PS3. As an example, my tyan based server with 2 X quad core 2.4 GHz Shanghai processors takes longer to perform a computation than my PS3. However, when I run the GPU based Folding@Home on my PC, its faster than my PS3. Hope that puts things in perspective for you.[/citation]You also have to understand what kind of WUs each one is crunching, X86 is crunching more complex WUs, while PS3 is crunching more complex ( for now) units then GPUs. Speed at the cost of flexibility.The PS3 is also highly limited to the platform, With its limited bandwidth and ram it will never come close to its potential.The estimated power your computer would consume over some unbelievable period of time would probably prevent it from ever cracking AES 128, nevermind 256. That and we are talking around millions of times the age of the universe if that age is around 13 billion years.So yea, you could program something to start a brute force attack. But then again you could use have an ante try and move a planet. Neither is going to have success assuming the passphrase isn't the weak link.[/citation]

Its too bad I never mentioned cracking AES.... other wise you have a good argument going ;) In fact the original article does not mention it either... No one is talking about cracking encryption here. Now if you would like to talk about cracking something like AES sure if you had unlimited funds and a 10,000 core super computer maybe in a few months you could do it...

Its all about guessing passwords not cracking some encryption. In fact at this point passwords are a far far weaker point then encryption, most people have trouble exceeding 8 characters for a password.


 

SlyNine

Distinguished
Feb 1, 2009
10
0
18,560
True, and I agree. A dictionary hack is the only reasonable way to go. My mind just seen PS3 used to do X, in this case catch child pornographers, and then seen it talking about encryption. After skimming through I skipped to the comments with most people referring to it breaking encryption ( or at least what I precived at the time). So my bad sorry for the wrong assumption.

You know, I love/hate the cell. Its a great little processing unit in its own right. But people think its some kind of supercomputer, credit to Sony marketing. But that annoys me as a Core I7 would crush it in most tasks, and a GPU probably in most other tasks. The Cell is a good middle ground though as it can do more kinds of work then a GPU can now, and of those highly parallel programs that the GPU can't do, it can do a lot faster then a normal CPU.
 
Status
Not open for further replies.