What's that rule? Our computing speed approximately doubles every year, and has for several years? Actually, I heard recently that our rate was increasing, up to more like every nine months, but for now let's assume we keep the current trend and double every year, like we have been doing for the last few decades. So 60 years, that's 2^60, or about a million, million, million times faster. (quintillion, for those keeping track) So, even if computing is a trillion times slower than projections indicate, they still do what you suggest.
Remember, even the first-gen cyberdecks were beyond supercomputers by our standards, the 6th-gen ones make those look like pong.

I'd argue with this. It's the same reason why your wiz-bang CD drive won't play DVD's. Encryption and compression methods, data formats, all sorts of transmission factors come into play. Granted, the carrier signal might be the same (and hence subject to jamming), but decryption would be a no-joy.

Except that any advance in computing power can also be applied to the encrypting. And encryption is easier than decryption. Massively easier. Linear increases in keysizes result in geometric increases in cracking difficulty. Switching from a 64-bit key to a 65-bit key will double the power required to crack the encryption in a given amount of time, with a trivial increase in power required for the encryption. Switching from a 64-bit key to a 128-bit key will make the encryption 2^64 times harder to crack. That'll swallow up all your hypothesized increase in computing power, and then some. And the best part... you won't even notice the difference in encryption times, even with a modern CPU.

distributed.net cracked RC5-64, using a distributed effort from hundreds of thousands of computers all over the globe. It took us five years. (And that only decrypted that data; it didn't break the algorithm. If we want to crack another instance of RC5-64-encrypted data, we have to do it all again.) If we assume that Moore's Law holds out for the next 60 years (it can't... Heisenburg will kick our butts long before then), a similar 2064 effort to break RC5-128 would take eighty years. And RC5-128 encryption can be performed on the fly almost invisibly by modern machines. Imagine the kind of encryption that machines 2^60 times faster can use...

You have to assume a qualitative improvement in decryption methods. Simply assuming a quantitative improvement in the amount of processing power you can throw at the problem puts you on the losing end of the battle.

So you can feely distribute the encrypted file with no worries, you only have to keep track of the key.
But...If the key is just as big as the original file, why not just keep track of the original?

p.s. This is a joke. I DO see the potential for increased security (needing both files), if anyone tries to explain it to me I've got a carp here with your name on it.

BTW, most of the powerful systems ARE small. Cyberdecks are like the size of a bible or small dictionary, IIRC.

How hard is it to simulate quantum computing? ie, without a real quantum computer? This is all pretty unlikely (as I imagine that emulating QC is pretty Hard as computing tasks go), but perhaps the SR tools use emulated quantum computing, and therefore the different ratings of encryption/decryption simply reflect a more robust and optimized emulation?

probalby not. Honestly, I think I agree totally that the encryption rules are completely knackered, and made a total sacrifice of real-worl-believability for game-world-balanceability.

That said, snooping encrypted communications or data STREAMS could be more possible -- poorly implemented encrypters could have a lot of redundant data, etc, or known plaintext portions, which would make cracking the data significantly easier in a data-analysis sense. Or they might have a [known?] weakness in their random key generators, or things like that. Maybe then the different rating decrypt software isn't reallt about true encryption-cracking, but about data-analysis to find keys and thereby uncover the data. Kindof like how l0pht-crack can pwn windows passwords, lol.

That said, it should be extremely fast and easy to encrypt something of nearly arbitrary strength, given the computing and media-processing power of the matrix (think of the massive parallell processing and transfer abilities needed for a sculpted host, for example), and still take a Long Time to crack.

I'd be interested in seeing how well decking would work if we fiddled the encryption-related rules to be more realistic. =) I like John Campbell's point -- it might make it harder to do thins the Current Way, but it could potentially open all sorts of other doors in terms of story options. On-site access would be more important, and legwork to find existing legit users of the system (to snoop their communications or swipe their hardware) would probably become more prominent.

Of course, this makes Otaku all the more powerful, I imagine, in that some of them are probably certainly encryption savants

Ah, but essense cost is not solely about Size, but about the degree of invasiveness for its' connections to the body. I bet that cranial cyberdecks are not simply a reduced size deck crammed into your skull, but uses portions of your wetware to aid in processing/connectivity.

Yeah, I figured the emulation was pretty infeasible.

re: encryption:
I considered it less Computers, and more a Knowledge skill. Why? Because cryptanalysis is much more about heavy theoretical mathematics than it is about computing. That's why the NSA is (I hear) the largest employer of mathematics majors.

This is all very true. It is also not encryption, according the the definition of encryption; none of it is. This is what Scramble IC and drek like that are, all of which I agree should be about as noticable to a true decker as a fart in the air conditioning And, unsurprisingly, it is; it's both cheaper and more effective to simply translate stuff into Klingon or something than use any form of "encryption" listed in the book, with the possible exception of rigger network encryption.

But there will still exist *real* encryption as well, stuff that is built to be secure, unbreakable, with open-source software that actually *encrypts* stuff. The fact is that encryption theory and implementation must be open to the public in order for it to be useful, and I expect at least one AAA megacorp will realize that. Even if this proved untrue, open-source software, such as stuff puclished under SourceForge or GNU, isn't going anywhere soon, and in a dystopia world like Shadowrun I imagine such a concept will be all over the underworld, especially in forums like Shadowland. And that *real* encryption should be available to runners as well, even if "the man" makes it Legality code ZZZ and you are made to disappear along with your friends, family, and everyone else in a six-block radius if you're ever caught using it.

It already is that way. Have you checked out the encryption rules for anything other than Matrix encryption lately? L3 Encryption is enough to regularly stymie every decrypter available, and L5 is secure against everything but EW-skills of 20+ -and- karma pools of 10+

~J

Are we reading the same SR3? Broadcast encryption is ironclad.

~J

I disagree.

If encryption were nigh-unbreakable (via bruteforce - an important distinction), then things would still be fairly balanced.

On one hand it makes some things HARDER for both the GM and the players. GMs need to provide avenuesfor players to get oast things that are encrypted. Players need to findways to do it. For example, rather than BAM breaking some communications, you snoop the traffic for a while, look for patterns, etc -- this could be a tine-based test where skill/utility grooviness would reduce the time needed to crack something. Cracking open an encrypted file archive, on the other hand, would be reaaaally hard...

The keys might be on the system - maybe the decker can find it. This is the weaker solution (and one often used today -- your PGP keys are stored locally, in an archive encrypted with a user password). If your decker finds a keystore (perhaps in the user home for the last person using it, or some commonly used place (you'd besurprised how few people change the default installation locations )), they should be able to try cracking that -- and since it's probably a user password, you could run like a dictionary-or-other crack on it; l0pht-crack is VERY good at this, for example. Most users are idiots, and won't use a strong phrase/icon/whatever. (Matrix passwords might be more tricky - who knows, maybe it's an icon you present or something ...) This could be the in-game interpretation of what cracking encryption on a file means in the current sense (roll , apply decrypt utility, etc).

The keys might be on a token or laptop of some employee or whoever needs access to it -- well, snoop those access logs, and see who has access / last used it! Now you know who to talk to or jack-some-goods from. (This is a common thing in the security industry for those who needsomethin secured -- smart cards or PCMCIA tokens are often used.) This would be for "more secure" things. This would open the doors for all sorts of other ways of getting an encryption key for Important Data than simply having your decker sit at a terminal for a few hours.

It might have been a one time pad, in which see the above case -- you need to find one copy of the pad that is used. This may require B&E, or kidnapping, or whatnot. *grins*

The key might be biometric - someone's retinal pattern or DNA as a basis for their user passwd, or something ... equivalent of having a token, but harder to change out if someone breaks/steals it.

I see this as allowing players and GMs to have a more "deep" world, with more "believable" consequences of actions, and more interesting ways of solving some solutions. For some things, the GM might say that a simple Decrypt test is all that's needed -- but for more Game-Important things (or that 300MP of research paydata that the decker stole while doing legwork ... ), the GM might impose more restrictions on how it has to be done. or incure a higher processing cost -- eg, encryption rating is so high that you'd need a mainframe cracking for a month or a year to hope to break it by brute force (which is still pretty unrealistic).

my 3

Ah-ha! I did not see this part. That makes things much more difficult, even at lower encryption ratings. See, I thought all you needed was one success to beat a Rating 10 broadcast encryption, which would happen roughly 25% of the time with a Rating 10 decrypt program, even without all the complimentary and other dice you add in. This makes it much much harder to crack. Objection withdrawn.

if you have a problem with that then stop playing the game called SR. yes its a game, not a simulation of real life

hmm, it seems that breaking data encryptions are easyer to do then breaking broadcast encryption right now as from where im reading data encryption dont have a required number of successes. in fact it looks like data decryption works more on the idea that any software have bugs, and if you know about these bugs you can get the software to behave outside of its planed ability. the decrypt utility is more like a "intelligent" exploit (it can not only use known ones but can probe for unknown ones somehow) then a bruteforce password/key tester (alltho it may do this to, dictionary tests and so on).

in fact it seems the entire decker attack scenario works on this idea. the toolbox pokes for flaws in the target code and the decker then supplyes the specific commands to get the ball rolling. we must allso not forget that while broadcast encryption have a max rating of 10 (look at hte table there is no 10+ entry) and broadcast decryption have a max rating of 8, the data encryption and decryption ratings both have 10+ entrys. therefor they can go as high as you want them to. and as you go higher the numbers gets slightly insane, and if your working on a host you have the added problem that any decryption test is a system test, therefor risking that your security tally goes up. so you better get it right the first time or else.