CircleID

Back in March, it was widely reported that RSA had suffered a serious security breach that (to some extent) weakened the security of its SecurID token. However, the NY Times reported then that the chairman said that the penetration wasn't absolute but "it could potentially reduce the effectiveness of the system in the face of a 'broader attack.'".

That broader attack may now have happened. Press reports say that Lockheed Martin was attacked, possibly by someone exploiting the RSA penetration.

This incident reveals the dangers of companies like RSA not being open about security incidents. Many companies (and many government agencies) have long relied on SecurID tokens. Without details about the problem, though, it is unclear how they should protect themselves. Get new tokens? Change employee PINs? Firewall off the administrative servers? We don't know — and that's the real problem.

(I confess that it isn't clear to me just what RSA is protecting by not revealing details of the danger. Its own reputation? That suffered a big hit in March. Its product sales? They might drop very sharply now, since it seems that even a sophisticated customer couldn't protect itself following the breach. What attack was enabled by the stolen data? If the RSA penetration really was an "advanced persistent threat", as they claimed at the time, the attackers certainly had the skills to discover that on their own even if they hadn't known it already.)

The really interesting question is what proper response is. Should companies be required to disclose problems that could adversely affect their customers? Are companies that do not make such disclosures civilly liable if harm could be prevented by timely disclosure? If they aren't liable, should they be? It is past time, I think, for such a discussion to take place.

By Steven Bellovin, Professor of Computer Science at Columbia University. Visit the blog maintained by Steven Bellovin here.

Related topics: Cyberattack, Policy & Regulation, Security