Websense has developed a method of sniffing out malware using Google’s binary search, a little known Google capability that digs deep into executable files.
Half of this is very interesting, and the other half scares the daylights out of me. The idea of being able to scan file internals, combined with Google’s crunching capabilities, makes a future Websense offering a heck of a proposition for realtime discovery of drive-by threats (and makes its acquisition by Google a possibility as well).
The fact that Google can crawl at this level, combined with desktop search, puts everyone’s archives at risk of leaking into the open, something I don’t like so much. I don’t use desktop search, but I suggest that anyone who does invest in some type of disk encryption for sensitive files. I don’t think Google would do anything bad here, but with a byzantine company you never know what might actually find its way through the cracks, and it is obvious that indexing isn’t limited to web pages and PDFs anymore.