Posted by: Arian Eigen Heald
Admins and Auditors, Data Breaches, Database, Database security, DataManagement, Security
The recent report on the Countrywide data theft got me thinking again about how to monitor insider access to databases.
The story is that the thief had access to the Countrywide (a mortgage broker) set of databases, which, of course, held all sorts of private financial information. A treasure trove, in fact, for anyone seeking a quick buck on the Internet. Countrywide is owned by Bank of America, and I have to wonder if they had done a third-party vendor review anytime recently, or had relied on Countrywide telling them everything was secure (which lots of Banks do, despite the Office of the Comptroller of the Currency telling them NOT to do that).
According to Countrywide, “The thief took advantage of a lapse in policy.” What interesting language. What policy, exactly, and what defines a “lapse?” Sounds like nobody was really paying attention to database access. Did Countrywide or Bank of America discover the thief? No, the FBI did.
What would it have taken to catch the thief in the act? Given that the thief was a “senior analyst,” it means that controls would have to be really specific. Let’ s brainstorm a little bit, because we need to start thinking this way. Too often, insider access is left wide open, and excuses are made that “it’s too time intensive” or “it takes resources away from the server.” Those excuses will no longer hold in court of law.
Think about it: if your organization offers up those excuses, the judge will jump all over you. If you have done background checks and are monitoring access, a lot of time and money will not go to lawyers. Demonstrating “due diligence” with regard to your employees and your data is very effective.
So, how could we monitor that kind of data? Two thoughts occur to me: first, only allow the employee to access records he works directly with, and require approval for access to any other records. This won’t rule out collusion, but it will make it harder for a single thief.
Second, log use of flash drives. This could be “silent” logging, but you could put two and two together, if the databases were also logging access.
How would YOU catch him?