At 06:36 PM 09/29/2001 -0400, Declan McCullagh wrote:

[Karl is on the ICANN board of directors. --DBM]

Yes, but he's got the "Hey, how did somebody from the *public* get one of the public slots on the board of directors?!?!?" seat on the board, and the cabal is trying to prevent anybody like that from actually happening again :-)

From: Karl Auerbach <karl@cavebear.com> To: Declan McCullagh <declan@well.com> cc: <rforno@infowarrior.org> Subject: Re: FC: Richard Forno on ICANN and Net-stability against terrorists ... I've spent a lot of time dealing with capability based operating systems (a technology that I believe deserves to be revived), mathematical expressions of security policy, formal proof of correctness of operating systems, real-live inplementations of secure operating systems and networks, cryptographic engines, key management systems, etc. Most of the document are buried deep in paper archives at the old National Bureau of Standards. As for the software and networks: Who know where they might be these many years later.

The current Capability-based operating system project is EROS, the Extremely Reliable Operating System, www.eros-os.org (I think it's .org. You can't just look for www.eros.com for obvious reasons:-) I know it was active a year ago; not sure what they've done lately, but the last I heard, their ringleader was at a university professorship, so there may be grad students developing it some more, and they were looking at making it microkernel-based. One of the other well-known capability-based systems was KeyKOS; some of the main folks from that are at Agorics.com.

And let's put things in perspective. What we're going to be doing is looking at many non-technical protections, like making sure that there are sufficient backups and procedures so that DNS infrastructure can be repaired.

There are three or four main kinds of attacks/failures for the domain system - - Failures/Attacks on the servers themselves - Attacks on the data transmission - various technical attacks such as spoofing DNS requests, adding extra records to responses, etc. DNSSEC and similar kinds of authentication are important for preventing these. - Attacks on the data, e.g. forged change requests - that's another very strong reason for authentication technology, whether it's PGP or dumb passwords. - Social engineering attacks on the system, such as UDRP abuse by trademark holders trying to reverse-cybersquat, as well as cybersquatting abuses, and delaying the extra TLDs. James Love <love@cptech.org> suggests:

Can we spell, mission creep?

Depends on what you think their original mission was :-) If your initial goal is World Domination, mission creep's no problem... I've been especially concerned with their grab of the IPv6 numberspace and the price-fixing that's prevented almost anyone from buying it.

From today's WSJ

http://interactive.wsj.com/articles/SB1001643073146154880.htm Some of these computers, such as the primary "A" root server in northern Virginia, operate within secure buildings, but others are far less protected. When congressional auditors recently checked the security surrounding them, "one of them was sitting in a professor's office at the University of Maryland," says Keith Rhodes of the General Accounting Office. "I would worry."

There was a while that MAE-East was in a parking garage in Maryland, with not much more than chain-link fence for protection.