Its a free country. Youre allowed to say all kinds of distasteful and offensive things, especially when youre telling the truth. And youre allowed to disclose security vulnerabilities in IT products. But theres still a right way and a wrong way to do it.
Michal Zalewski has problems with Microsoft. He thinks Microsoft doesnt cooperate with the security research community, that in fact it abuses that community, that it disserves its customers and doesnt take security seriously.
For all this, when he discovered a bug in Internet Explorer that caused it to crash, he decided not to disclose it to Microsoft through the normal channels ([email protected]) but instead to disclose it publicly on the popular Full-Disclosure mailing list.
Its important to note at this point that, for all the use of terms like “highly critical,” theres still no actual claims by anyone that they have been able to exploit the crash in order to execute attack code. Zalewski is very clear about this himself, and much more reasonable in his evaluation of the severity of the vulnerability than some outside agencies like Secunia.
Many crashes like this turn out to be exploitable, but not all of them do, and some are only exploitable with great difficulty. In other words, this may turn out to be nothing more than a way for a Web page to crash your browser.
Or it might turn out to be exploitable, in which case its a way for a Web page to execute arbitrary code on your browser, including installing programs.
Lets assume what is likely, that Microsoft found out about this vulnerability the same time as everyone else, when the company read it in Full-Disclosure. That means that Microsoft has only had since then to research it and work on a patch.
Lots of other researchers dont like Microsoft but still give the company first shot at vulnerabilities in its products. Some hold off however long Microsoft takes, some give the company a specific period of time. Consider this example that just came cross the wire—Matthew Murphy gave Microsoft six months, during which the company partly addressed the problem.
But Zalewski, as a protest basically against how he says Microsoft treats researchers like him, decided to just disclose. It seems to me that Zalewski thinks that its researchers who matter, but in fact its users who matter, and his action is obviously harmful to users because it increases the likelihood that an exploit will become available before a solution and it increases the duration of that period of time. Just who does he think he is?
Actually, Zalewski says that he does things like this “whenever I can reasonably believe that no immediate harm would be done to third parties.” This seems inconsistent to me. If he really does think that this might be an exploitable vulnerability and he discloses it without advance notice to the vendor, of course theres a chance of harm to third parties. He just wont let that get in the way of him showing off what a troublemaker he is. Wow, Im impressed.
Its frustrating to see Microsoft take many months at times addressing serious vulnerabilities, but there are times when it absolutely makes sense. One of the most famous examples is MS04-007, the ASN.1 vulnerability. It was an extremely serious vulnerability disclosed confidentially to Microsoft about six months beforehand.
The ASN.1 code was all in a single DLL, but its a widely used technology, and the test matrix for fixes to it was immense. According to Microsoft they found many problems in testing the various solutions, and every time they come up with another solution its back to the testing drawing board for multiple versions of Windows in 26 languages. And for some technologies (ASN.1 is probably a good example) you cant automate all the tests. Some of them need a human looking at the screen.
I dont know what the “borderline extortion practices” he refers to are, but theres no conceivable value to the public in him disclosing publicly with no advance notice to the vendor. With a serious bug this is on par with leaving gasoline and matches around and pointing out that there are flammable buildings about.
Zalewski may think hes some sort of hero disclosing this information, but his is the act of a vandal. If it turns out that the bug is exploitable and abused before its patched, then perhaps hell be proud to be remembered for that. The best we can hope from it is not changes in Microsofts behavior, but that his bad example will deter others from doing the same.
Security Center Editor Larry Seltzer has worked in and written about the computer industry since 1983.
More from Larry Seltzer