Interesting rumors flying around work...
I wish I had ran today. My heel is still a bit sore, but mostly today I avoided running because of laziness.
You can't really appreciate SCSI until you do a side by side install next to an IDE drive. My PCs will both be running SuSE 8.2 Professional with VNC so that I can control them remotely. I am also planning on installing VM ware with a version of windows 2000 on them... just for fun. Need a stable environment for experimenting with grid applications.
I have been studying reputation systems for the last month. Interesting that all of the so called GRID experts have missed the boat. This is the only way that dynamic grids will be able to protect themselves from code replacement attacks. Not to mention flooding attacks and other nasty things. These problems are incredibly difficult to handle while trusting the virtual organizations and trying to do user level authentication. None of the nodes have any knowledge. They can't tell whether a node communicating with them is evil or good. They can only see that "yes, this is a valid certificate". That isn't enough. Throw in a reputation system built on top of X.509 attribute certificates issued by distributed sources and you have something. Link this with a way of permanently registering nodes via their MAC address (so that they can't change their identities) and you have something more substantial than the toys that professors are playing with these days; toys that the industry seems to think are worth something, but then the industry has the resources to throw away man hours molding crap into something half-usable (at least the big players do). But that's perfect. The code was written for the big players anyway, so why shouldn't they suffer to make it work.
