Distributed computing is a somewhat unusual, and in some cases probably a questionable, pastime. But it may, in fact, be one of those cult sports that can suddenly creep up when no one is looking and become popular.

However, the contributing members of distributed.net take their sport very seriously.

Project Bovine, or RC5, is distributed.net's response to a challenge by the RSA to have an encrypted block of data decoded without the decryption key. Due to the size of the key, and the encryption used, it wouldn't be possible to find the correct key in a reasonable amount of time. Even with that brand new Pentium Pro it could take years!

To get around this kind of problem distributed.net have designed a distributed processing system which breaks up this huge job into bite size chunks which can be easily processed by special client software. Since there are thousands of distributed.net followers running a client on their machines, the combined computing power means that the result is within reach.

This concept raised some discussion at my work place over a number of lunch hours, as I had been making my contribution to project Bovine for the previous week or two.

After a little persistence, and someone pointing out that we didn't really have a good reason to try this out, we decided that the 300 or so Pentium PCs sitting in empty classrooms at our campus, were about to get a workout. The DES III challenge (a similar contest to RC5 but only scheduled to run for 3 days) was coming up and we were going to see how much we could contribute.

Upon first inspection of the client download page you get, what can only be described as, a sense of amazement. There is a client for just about everything - the most exotic being for Netware, RiscOS and OS/390. After a small debate we decided to go with the Windows CLI client. That way we were able to install and run the client via a Netware login script.

So that was the clients sorted out, but since all of our classroom machines are in the 10.18.*.* range of IP numbers I had to investigate the Bovine Personal Proxy server.

In the end I chose the FreeBSD version for two reasons:

1. My FreeBSD box has proven to be one of the most stable boxes in our server room.
2. 2) That box was the only non-production server already operational.

After the proxy server was up and running I configured the master copy of the client with the correct contact details and set up email logging. I then wrote a simple DOS batch file, which copied the client onto the local hard disk and then modified the workstation login script to run both the batch file and the client. All that was needed now was to boot up 300 PC's.

For what seemed like hours (but really only ended up being 30 minutes) we went around booting every classroom PC on our network.

We ran the client for about 36 hours, which allowed about 22 hours of DES III work and 14 hours of RC5 work. My FreeBSD server performed quite well considering that the proxy software wasn't very stable. Fortunately I was on hand to monitor it and restart the proxy process when necessary.

During that 36 hour period we processed approximately 9087 DES blocks and 6486 RC5 blocks. Our best rate was around 850 million keys per second. The FreeBSD server received over 2500 log emails of about 2k in length. During the period after the DES contest we ranked 11th in the RC5 contest.

It's worth noting that the FreeBSD server is a 486 DX4-100 with 12 MB of ram, 128Kb cache, 420Mb IDE hard drive and a 3COM 3C509 ethernet card. The OS version is 2.2.5-RELEASE and during the contest it ran a Bovine client, as well as pppd for my modem connection, constantly.

Unfortunately we did not find the winning key but it was a great learning experience. My only regret is not having enough time to set up the lab full of Power Macintosh computers to join the effort. Of course, without help from Alan & Marten it never could have happened. Thanks chaps.

Who knows? Maybe next holidays we can get our other campuses on our side and triple our computing power!

Related URLs:

- Joel

Return to Issue #2