This week Microsoft announced important updates to policies around discovering and disclosing third-party software application vulnerabilities. They’ve officially expanded their Coordinated Vulnerability Disclosure (CVD) policy (launched last summer as a replacement/renaming of their “responsible disclosure” policy) and have made public an internal employee policy (launched in November 2010), which requires in-house researchers to adhere to CVD guidelines, and report vulnerabilities in third-party products to the Microsoft Vulnerability Research (MSVR) program. MSVR then reports the vulnerability privately to the vendor and coordinates with the vendor on its investigation progress . In a related gesture, they released inaugural MSVR Advisories on vulnerabilities discovered by Microsoft employees in Chrome and Opera (fixed by the vendors in the latter part of 2010).
For more background, here’s some of what Microsoft has said about their updated vulnerability research policies:
My first comment is that Microsoft should without a doubt be commended for taking such an active role in protecting their customers not just from security weaknesses in their own technologies but in third-party software also. Microsoft is trying to fill a void that has been created in the vulnerability research space — the gap between researchers discovering vulnerabilities and actually reporting them back to the software vendors.
While Microsoft is looking to fill this void by doing vulnerability research themselves they, and other technology companies, should look to solve the two larger problems of why vulnerability researchers have in large part abandoned working with vendors.
1. Money on the table: Vulnerability research is not easy work and researchers now have an outlet to be compensated for their work by selling zero-day vulnerabilities, both to good and bad intentioned buyers.
2. Mistrust of vendor accountability: Vulnerability researchers who are less motivated by money are still extremely dissatisfied with the time it takes for vendors to fix vulnerabilities reported to them. Also, there’s a genuine sense of resentment among researches because of games sometimes played by vendors. Microsoft, and other technology companies, still fail to set a timeline during which researchers need to wait for Microsoft to create a patch, but after which a researcher should be able to publish details to help the community without being vilified by Microsoft or other technology companies.
Is Microsoft doing vulnerability research going to help their customers? Most definitely. But not as much as they would help customers by finding a way to compensate researchers and stick to a measurable time period to produce a patch. There is no comparison to the exponential benefit Microsoft would have on product security by bridging the gap with the research community. The community will always be stronger than any in-house Microsoft efforts at vulnerability research and that right now equates to more zero-day being found in the wild.