Earlier today, I was in my divisional assistant director's office with my boss discussing NC public records law as it relates to email, web usage, documents and etc... As you may or may not know, I work for a local government, and the proportional chunk of our budget that we spend to retain public records is non-trivial and mandatory.
Even so, we face many challenges. The challenge i'm hoping someone at google can address for me today relates to classification of documents and integration with 3rd party solutions.
Some background:
Even though a large portion of what we do is public records, a non-trivial portion of it is by no means public records, and in fact we have a legal obligation to protect some of the data from being accessed outside of whatever use it is intended for. Personnel records are considered proprietary, as is information we have concerning direct deposit for employees, etc... Things that a lot of lawsuits would be filed over should they fall into the wrong hands.
The problem for us is that unmanaged document sprawl has both public records and private information scattered across terabytes of storage in a forest of directories, complicating our compliance efforts to no end. Also, not everything that is not private information needs to be retained per public records law, and so can be deleted to lower storage costs once the information is no longer relevant. Imagine signup sheets for 10+ years of employee birthday parties, directories misused for music storage, pictures of someone's kid, etc...
We would also like to explore the possibility of DLP to protect what needs protecting at some point in the future, as part of a larger and ongoing risk management and security management process.
I proposed automating the classification process with a sort of google appliance, or the appliance of a search competitor (dont know much in the document management space), and once we have a handle on that, using that appliance to provide data classification for a DLP solution down the road. In my mind, I can see the perfect solution to all those challenges on one box, but I want to check in with google and the internet to see if it's possible, been done, or being done.