2005-05-10 22:26
by Vasil KolevIt’s official, Larry Lessig is coming to Sofia, here are the details and the agenda. We’ll be recording him :)
Tomorrow Bulgara will be performing in O!Shipka, come to hear them :)
Because the lists that I sent last night UniBG test’s description don’t have a public archive (even users@unibg.org), I’ll describe here what happened.
The test consisted of connecting a server with NICKLEN=2 and seeing it’s effect on the network. If the server is hyb6, it doesn’t matter, but hyb7 starts killing all the users with longer nicks, using it’s full bandwidth. The result was murderous, from 20000 users the count of users fell to ~1800 in 10 minutes, and would’ve probably fallen to ~100 (constantly leaving and connecting) if the test had continued for 10-15 minutes more. A side effect was that services, sanserv and the other similar stuff in the network was unable to keep up and fell (some servers got disconnected, too).
The bad thing is that it’s extremely hard to stop such problem, if it had enough time to kill most of the network – in the end the server that was used to make test was juped after I stopped it. The normal way of stopping something like this includes firewalling the offending server from the network or touching the config of the servers it connects to. The threat is not that minimal – the history knows of at least 3-4 intrusions in irc server in UniBG.
About the other complaints – I warned before the test, and made it in not-so-peak period.
I also saw the history of UniBG, that’s on cult.bg – stupid stuff, damn those pseudo-historians, this one is like a tabloid journalist, he described only the scandals, exaggerated a lot, wrote some pretty fake stuff, etc… In the end I’ll collect an archive of the lists from 97 to this moment and will publish it, for the people to be able to see what’s true and what – not.