Why you should block Kazaa and encourage better P2P.
At a recent demonstration of a network statistics tool, the presenter pointed out a bar on a chart and said that about half of the bandwidth used at the site was for a peer-to-peer (P2P) system called Kazaa.
P2P systems have enormous potential to free independent artists and Web sites from big ISP bills, sneak political information past censors and, perhaps most importantly for our readers, provide a challenging new platform to develop software ideas. But, Kazaa is proprietary, mostly unencrypted and, with a little work, can be filtered out. On page 56, Chris Lowth explains how.
Please do it. Read the article and filter your company's Kazaa users into oblivion. It's one of the projects you shouldn't have trouble convincing management to approve, because legal threats for hosting illegal copies of files on P2P systems are big news. With the techniques covered in Chris' article, you might be able to get a Linux firewall approved for a company that doesn't use Linux yet. There's always a first Linux project for every company, and this makes a good one.
However, there's another, more important reason to block Kazaa. The gap between what P2P needs to do in order to be a useful free and anonymous speech system and what Kazaa bothers to do, is shocking. If you can detect Kazaa traffic, repressive regimes certainly can. Help encourage the development of strong P2P by blocking inadequate P2P.
Ready to start working on a P2P system of your own? Brandon Wiley introduces a fundamental technique, distributed hash tables, on page 71.
And, because the theme of this issue is system administration, we cover two of the classic problems: how to design a backup plan for quick and painless restores and how to set up robust, sane file servers. Now that blank CDs are cheap, it makes sense to give each system its own custom recovery CD. Craig Swanson and Matt Lung cover Mondo for bare-metal restores on page 46. Erez Zadok lays out the problems with NFS and covers a flexible automount solution you can live with, on page 52.
After the movie industry, the next big leap for the Linux desktop is into electronic design automation (EDA). Not only are open-source EDA projects gaining features, the established EDA vendors are porting their proprietary tools to Linux as well. Michael Baxter explains how you can do hardware design almost at the speed of software with Xilinx chips and development tools on page 74.
Reuven Lerner's series on open-source content management systems for the Web continues with coverage of Bricolage on page 12. And, Eric Jeschke is back by popular demand with more GIMP techniques on page 80.
All you kernel hackers and scalability fans out there can check out page 18 where Paul McKenney explains the read-copy update (RCU) algorithm he developed at IBM and how he and other kernel hackers are making it work in the eagerly anticipated 2.6 kernel. Don't forget to download and test a 2.6 kernel with your own favorite apps and hardware. With the freedom to hack comes the responsibility to report bugs.
