MS knew of Aurora exploit four months before Google attacks
China light on the matter
Microsoft first knew of the bug used in the infamous Operation Aurora IE exploits as long ago as August, four months before the vulnerability was used in exploits against Google and other hi-tech firms in December, it has emerged.
Redmond's security gnomes finally got around to patching the exploit on Thursday. Microsoft's advisory accompanying its cumulative update for IE credited Meron Sellem of Israeli firm BugSec for reporting the HTML Object Memory Corruption Vulnerability (CVE-2010-0249), the zero-day vulnerability used in the now infamous attacks.
BugSec's bulletin states that it reported the bug to the software giant on 26 August. The bug affected IE 6, IE 7 and IE 8 (the latest version), but the hack attacks against Google et al targeted IE 6, a browser first released in 2001. Exploits involved tricking users of vulnerable browsers into visiting booby-trapped websites. These sites downloaded the Hydraq backdoor Trojan and other malicious components onto compromised PCs.
ThreatPost, a Kaspersky Labs news service, reports that a patch against the flaw was lined up for release in February. It was published early in response to the row that followed Google's surprise admission last week, that the bug was being exploited in cyber-espionage attacks targeting it and other hi-tech firms.
Software vendors in general often take months to develop security fixes, a process that often involves a great deal of testing work. An unfortunate set of events meant that this particular bug became one of the most infamous in years. Even now we know the details, the future potency of the bug is far from immediately apparent.
A quick search of Secunia's database, via its PSI patching tool, reveals a problem with an unpatched ActiveX control that looks just as bad, for example.
More discussion on whether Microsoft's patch was tardy or not, and the role of the vulnerability in the Operation Aurora attacks (it may not have been the only vector), can be found in a blog entry by Graham Cluley of Sophos here. ®
it's shit like this that makes the "full disclosure of flaws" camp look ever so right in their attitude. When Microsoft hides flaws, everyone suffers. Except, it seems, Microsoft.
Yes, please, disclosure++
It's not just Microsoft, a lot of outfits with closed-source software will leave horrid exploitable holes open for months, sometimes even years. Apple are as bad, scarily.
I can see the logic of the argument about keeping things quiet until it's fixed- make no mistake about it. However, the argument only holds up when the developers actually fix the bugs in a responsible time frame.
Personally, I am in favour of full disclosure- and I also prefer completely open software, especially for security-critical stuff.
However, if you must be closed, and if you can't support full disclosure, maybe disclose to the vendor in private with a notification that you'll go private after a month. If the software concerned is so vital and sensitive that it's too special for initial full disclosure, it's important enough to put full resources into fixing and testing in that (non-mythical man) month.
If you can't be arsed to devote resources to fixing your shit rather than working on new shiny things for marketing to push, then screw you- you don't deserve the month's head start, and the deference.
The guise of first reports.
The guise of first reports always linger longer as they tend to rannk higher in peoples mind longer even though information may have changed.
Thus knowing the above I am sure lots of people still think China is the guilty party for hacking when in fact the code was brewed in the USA.