Tweet

At the end of May, two Google security engineers announced Mountain View’s new policy regarding zero-day bugs and disclosure. They strongly suggested that information about zero-day exploits currently in the wild should be released no more than seven days after the vendor has been notified. Ideally, the notification or patch should come from the vendor, but they also indicated that researchers should release the details themselves if the vendor was not forthcoming.

This is a pretty aggressive goal. Microsoft, for example, does not set public deadlines for critical patches: a balance needs to be found between quality and speed. Balancing the two is not that easy. On one hand, an actively exploited vulnerability will be used by malware writers aware of the vendor’s vulnerability. On the other hand, a quick patch could have negative side effects and could cripple the application or the entire system.

Almost every security vendor has false positives that result in negative side effects. We have lots of safeguards in place – such as checking against whitelists – but Murphy’s law applies. Our industry has crippled the computers of users. What more an operating system vendor – they need to be extra careful in patching. They need to conduct proper quality assurance, as the patch will affect millions of computers.

In my opinion, disclosure after 7 days is reasonable and OK, but expecting a patch in this time frame is unreasonable. Let’s see how Google itself will manage. Currently, there’s a Google Android Trojan spreading which is able to hide itself from the “Device Administrator”, which renders it invisible from security programs and clean up attempts. This was possible because of a security flaw in Android. Will Google be able to fix this within 7 days? Let’s see…

The bigger debate is not just about how long it should take to report and fix vulnerabilities. It’s about how vulnerabilities should be reported in the first place. If you believe a recent media report, the US government is now the biggest purchaser of malware. How do we ensure that the affected vendors are informed, that these are not used for offensive uses like Stuxnet? How do we ensure that these same vulnerabilities don’t end up in the hands of the underground, which will use these threats widely?

What needs to take place is a bigger discussion around how discovered vulnerabilities are dealt with at all. These should be between all those involved in this field – developers, governments, and researchers – to determine how we can deal with security vulnerabilities in the future.