[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Anti Nick Collide...



On Jul 26, Matthew Sullivan wrote:
| TS seems to be implemented in EFnet, but not IRCnet, i would assume that it could require a great deal of server resources for the TS on nick, also it would not be fool proof against attack.

right, see http://www.stealth.net/~kalt/irc/TS-Delay.html
and http://www.stealth.net/~kalt/irc/stats/collisions.html

| My Suggestion:
| 
| Rather than TS on nicks, have a method of modifying the nicks (forced) to avoid nick collisions.  This has been comented on by others in that the server could have it's own identifyer ad asingle char to the last char of the nick, either changing the last char or adding an extra char to it.  My first thoughts was to generate 2 random nicks and force the change, however in the event of mass collisions i would conceed that this would more than likely create a larger problem of CPU resources.  
| 
| Perhaps a simple solution of generating the last char could be based on current unix-time and quick calc to convert to char.  Would this require overhead?  Is there a better idea?

this isn't really new and has been discarded before for
several reasons.  In short, it is kludgy, and there are
better ways to solve the problem (but harder which is why it
is still to be done).