title:

Cyber war and cyber isolationism

By Scott Bradner

There has been a bit of a splash in the press the last few days about a mention by ex CIA directory Michael Hayden of the idea of creating new, extra secure Internets for government or commerce. Users would have to give up their privacy to use these versions of the Internet, with a requirement for the use of real names and all their traffic subject to deep packet inspection. The vision seems to be that government would use one such network and services such as banking would use another. This is not a new, and not a good, idea.

At least some of the talk comes from a very interesting symposium at the Potomac Institute for Policy Studies on "Cyber Deterrence: Mutual Assured Disruption or Other Options." (http://www.potomacinstitute.org/index.php?option=com_content&view=article&id=985:gen-hayden-to-headline-next-cyber-challenge-symposium&catid=65:past-events&Itemid=94) The symposium included Gen. Michael Hayden, USAF (Ret.), currently a Principal at the Chertoff Group and a former Director of the CIA and NSA; Michael Tiffany, Chief Architect at Recursion Ventures; and Dr. James Mulvenon, Vice President, Intelligence Division, Defense Group, Inc., all of whom presented provocative topics for consideration. This was a very full symposium and the mention of separate secure Internets was a very small part of it. (for example, starting about 89 minutes into the video and again at about 109 minutes)

The idea that a separate or at least filtered network will magically fix security is not new. There was a short lived boomlet for a govnet (a private network for government) back at the beginning of the century. (see http://www.networkworld.com/news/2001/1128privatenet.html and http://www.networkworld.com/supp/government2002/fts2001side.html and "Does going it alone make sense?" - http://www.sobco.com/nww/2001/bradner-2001-10-22.html.) Such proposals turn out to not do much in the way of real protection, among other reasons is the vulnerability of the end systems. If your machine is hacked then any advanced security network that uses identity credentials from your machine for access control will be vulnerable.

Michael Tiffany pointed out that the vulnerability of the end systems, among other things, mean that the all too common assumption that all security problems would be solved if we just eliminated all anonymity on the Internet and were able to attribute all traffic to someone. (about 97 minutes)

There was also a bunch of discussion about the effectiveness of

cyber counter attack as a deterrent - something that the U.S. Deputy Defense Secretary William Lynn is likely to expound on during a speech on July 14. (http://www.sfgate.com/cgi-bin/article.cgi?f=/g/a/2011/07/10/bloomberg1376-LO6BDO6KLVRI01-3UVV6RDOF9RPOHK3N55MK10LPA.DTL) - the clear message was that it is unlikely to be a good deterrent because of the difficulty of figuring out who the actual attackers are and who might be directing them.

The most important conundrum facing those who want to solve the cybersecurity problem was best described by General Hayden. He noted that the Internet was now the basic communications mechanism of everyone, but that history has shown a need to keep the government weak when it comes to controlling basic communications mechanisms. This conundrum makes things hard if one believes that government is part of the answer but not everyone thinks that government is the best protector of individual liberties.

The panel does not paint a rosy picture, nor should it have -- there are too many easy answers, such as govnet, floating around , where he hard answer that would actually make a difference - harden the end node - is given short shrift.

disclaimer: Harvard has classes in the art department that discuss rosy pictures but the university has not expressed an opinion on the content of the Potomac Institute symposium so the above video review is my own.