It was nice to see Marcus Ranum (who has an interesting slant to the security industry) get some press again. This time it’s on responsible / full / no disclosure. In a probably unrelated attack, his site is defaced by a SEO blackhat. Irony, eh? If only he had patched or used software which has learnt the hard lessons.
Here’s the anti-rant I wrote my co-workers a Friday or two ago:
Ranum’s argument has four major elephant sized flaws (at least).
Firstly, he states that security has not gotten better. This is clearly wrong. Security has gotten a great deal better, but so have the attacks and our knowledge. However, the impact of attacks has been steadily decreasing. When I first joined the Internet, there were perhaps 100,000 people on it at a very small number of sites. That year, the Morris worm nearly destroyed the entire Internet. There have been no significant attacks like that for some time. Yes, there are more attacks, but considering there are more than a billion of us on it now, that’s to be expected. Attacks require a great deal more skill today than in Morris’ time. Old software, particularly in the webappsec is trivial to exploit. Proof – modern stuff which is hardened through the lessons we’ve learnt is very hard to exploit. Software which does not heed the lessons is trivial to exploit (see MJR’s site, natch!). Without some pressure, all software would be trivial to exploit, not just the lesser used stuff.
Secondly, he states that disclosing vulnerabilities is akin to shouting fire when there is barely any smoke. The implication is that you should never shout fire, even if there is the possibility of fire. However, if no one shouted fire, children’s pajamas would still be made of highly flammable materials resulting in third degree burns or death instead of slow or insulating materials we have today. Only through research, standards and indeed, advocates (akin to vulnerability researchers) doing shock stories on tabloid TV did we move from obviously deadly dangerous to moderately safe. Fire is a particularly weak analogy as the metaphor breaks down very quickly – fire always occurs and is a natural phenomena.
Thirdly, Ranum ignores evidence that contradicts his position. Vendors and customers are hurt by rampant full disclosure, and I agree that some folks are only out to get on CNN for a few cycles. However, responsible disclosure is the only proven way to make security sloppy companies like Oracle pay attention – eventually. It made Microsoft more secure, and I think if you look at NT 4.0 (1996) versus Vista (2006), Vista is a much larger but harder target. Oracle’s CSO (is in my view) negligent because she thinks like Ranum, and refused to protect her customers and ipso facto all of us.
Lastly, Ranum HATES – and I mean truly despises – upgrading software. This leads directly to his point of view that if there was no disclosure, there would be no (or much less) patching, therefore he wouldn’t have to upgrade. This is a logical fallacy as one does not lead to the other. If all of us had his world view, we’d be running NCSA web server with no firewall on SunOS 4.1, i.e. completely unsafe. How would have Microsoft|Apple|Sun learnt how to secure (as best they are able) their operating systems without the challenges of security researchers and malware creators? It’s like MSRA golden staph – damn near unkillable around hospitals today. It didn’t get like that because we used soapy water.
He rants against the creation and sale of malware as if we’re powerless to stop it. However, it is already illegal to do this in many countries. So if someone writes malware, they are already breaking the law. Why would they stop now, or in the past in his alternate no disclosure universe.
I remember a few years ago that CERT sat on a major DNS issue for oh 8 years (I’m making this number up, but it was not a few months) until the last root server was upgraded to bind 8.something. There was an architectural flaw that could have destroyed the internet with a few packets. And I knew about this in like 1992 or 1993 and at that stage I was not in the security game fully – just a sysadmin. It only required someone with bad intentions and the Internet would have been dead. Why X years? Because there was no impetus to upgrade the root servers, despite it being 14 times redundant, simply because CERT sat on the problem. When I met Spaf a few years later at a SAGE-AU conference, I asked him about this, and he was unapologetic about it. Who gives him the right to decide if the Internet stays alive or not? It should have been fixed, and indeed it was fixed – eventually.
Will we ever be secure? No. Will Ranum’s or my site be safe from attack? Doubtful. Ranum is simply wrong in his thinking if by stopping disclosure we will suddenly become safe.
Ranum’s alternative is no alternative.
ps. I am no apologist for unrepentant full disclosure types out for their 15 minutes on CNN. Hint: I will never employ or recommend ANY full disclosure folks.
Leave a Reply