During a break in FSE yesterday Ross Anderson offered comments on why crypto could not protect privacy against weak comprehensive systems security. Much of what he said was confirmed by the news report today on new Dutch privacy invasion law for intelligence and law enforcement. He stated that traffic analysis is surely the greatest threat to privacy, and that there are programs in use by law enforcement and intel agencies -- with more advanced versions in the works -- that can log and analyze TA for global communications, thus greatly reducsing the need to process the full floor of communications that the NSA complains of being unable to do. >From that TA the agencies can identify targets for black bag job. And that crypto -- even end to end -- will not protect against bugs surreptitiously planted in those locations pinpointed via TA for acquiring vox and plaintext data outside the crypto loop. Ross said TEMPEST threats should not be overrated, and do not pose anywhere near the threat of TA and black jobs. I called his attention to the CRISIS report which recommended widespread use of strong encryption coupled with increased funding for other, unnammed, technologies for law enforcement. And that TA and black bag jobs had been mentioned as candidates for those unnamed technologies. We swapped tales about the weakness of bricks and mortar architecture against black jobs. And he noted that systems security is especially vulnerable to architectural attack, more so than many engineers trained in electronics are aware. Recall that the CIA's Special Collections Service performs black bag jobs which have been identified as needed by NSA intercepts, especially where encryption is used: http://nsa-cia-scs.htm The other day we transcribed a 1979 set of NSA specifications for SCI facilities such as vaults, secure working areas and entire buildings, which include a fair amount of systems security recommendations for architectural vulnerabilites -- various types of alarms, building construction, guards and response times -- as well as electronic sensors for equipment. http://cryptome.org/nsa-scif.htm To supplement that there's Willis Ware's classic 1967 study for the Defense Science Board on computer security: http://cryptome.org/sccs.htm In another indication of evolving policy on intrustive technoloy Willis says that he is not being re-appointed to the chair of CSSPAB, that new blood is being called for by the go-go folks pushing closer working relations with government and commerce at home and around the globe. He, too, wonders what will happend to privacy as invasive technologies move from natsec labs to the commercial world of suck data made so convenient by computer and network technology. Whether any of this will be fundamentally challenged by the House hearings tomorrow and upcoming on legal authorities of NSA and the intel communicy, remains to be seen.
