Confessions Of an RC5 Co-Conspirator: A Virtual Diary
Jason Compton

- Someone on IRC asked me today if I wanted to help out in cracking RC5. I have no idea what he meant. He tried to explain it to me. It sort of sounds like people are working on cracking encryption methods--probably because they don't like the American government's attitude towards encryption system importation and exportation. Whatever, sounds complicated.


- This RC5 thing came up again. More people are talking about it--I'm gathering that it's nothing illegal, it's some sort of sponsored contest. But working on cracking an encryption system sounds a little complex. There are massive massive tomes on cryptography--my roommate even has one. "Applied Cryptography." It competes with some phone books I've seen.


- Now these RC5 guys want me to talk about their project in the magazine. Ok, so now I've visited a website or two (http://rc5.distributed.net) (http://homepage.cistron.nl/~ttavoly/rc5/) and have found out that it doesn't really require you know anything about anything. See, the idea is that there are all these encrypted keys out there: 2^56 of them, in fact. You can run a client on your computer to try to find the secret key. If you do that, everybody on the team who was working with you gets pizza money or something like that. But I don't know. I'm a little leery of just running some net client because someone told me so...


- Ok, these RC5 guys won't leave me alone. I've been asked no less than 24 times to run the client and do an article about it in AR. Nobody seems to have had their system invaded as a result of the client, and the whole thing does look pretty legit. So, all right, all right already. I'll follow the procedure on the web site, benchmark my machine for their statistics, and go ahead and start running...


- Now that I've hacked a few code blocks...well, not really hacked them personally, but now that I've let my machine's CPU time be hogged a bit by this RC5 client, maybe I'll look into this a bit more. I'm a little disappointed that this is the "Amiga" RC5 effort but in the name of advancing the cause (and getting the Amiga name moved higher and higher on the list of top code hacking teams), non-Amiga machines are beig liberally employed in the effort. You can run these clients on PCs and high-end Unix workstations, too...and let's just say that most of those machines are pulling down more KKeys per second than my 060 card is.

Ugh, now they've got me talking technical. "KKeys per second" is the rate at which your computer is testing the keys to try to find the secret one. "KKeys" is short for "kilokeys", in other words, how many thousands of keys per second. My machine does about 100 KKeys/s: 100,000 keys per second. Sounds like a lot, right? First think about how much 2^56 is, then 100,000 per second doesn't sound very good anymore...


- Well, this is it. I've got the RC5 client running in a shell in the background. It's about 10% through a new block of keys, and it really isn't taking up CPU time--I can type here, carry on in IRC, browse the web, etc. I need to remember to break out of it with a control-C before I turn off the computer though, just to make sure all the hard work is saved.

So, if you spend a lot of time aimlessly online, or might want to start, putting your Amiga to work on the RC5 project is a good way to feel a bit productive. And the guys who run the Amiga RC5 effort will be happy--happy that I wrote this article, and happy that by the time you read it, the Amiga team might have broken into the top 10 teams in the international effort. It's an impressive feat--a ragtag group of Amiga junkies and guys with access to faster machines with a soft spot for Amigas have banded together to form one of the largest volunteer armies of spare CPU cycles in the world. I imagine you could make a movie like Toy Story this way, too. But it would probably be harder to motivate people.


Back to Features