Securing our databases and preventing the release of information our organizations have collected is an ongoing job for many of us. We patch our systems, ensure that our logins and users are granted just a few rights, encrypt data and backups, perhaps have various monitoring solutions, and certainly deal with plenty of stress at times. We work with network staff to ensure firewalls protect our systems. It may be a regular part of our job to actually argue with others that security is important for databases. If we work with regulated data, such as financial or medical information, then we may even have the struggles of compliance with auditors or even regulators.
All of our work can be for naught with a few simple mistakes that someone in our organization might make. Perhaps an employee takes a copy of production data on a laptop home and loses it. An employee might use the same password for a secure portal that they do for Facebook or some other social site, getting hacked and exposing our systems to others. We might even have an employee click on a phishing email or insert a random USB drive into their laptop and compromise entire infrastructures. As a security professional once noted, we have to win every time. The bad guys only have to win once.
It can be distressing, and even more so this week as I read this piece about medical data and Frank Abagnale, the inspiration for Catch My If You Can. In it, Mr. Abagnale states that he doesn't think technology will ever defeat social engineering, which is distressing to me. He may be right, though I certainly hope that machine learning and other technologies, along with lots and lots of data, will find ways to catch abnormal queries and data extraction, which often are a signal that potential data loss may be under way.
What's more distressing in the piece for me is the fact that some of this data, like the birthdays and SSNs, are stored for years. Unlike credit card data, which is more valuable right away, unchangeable data becomes more valuable over time, so it pays to keep it around. Maybe the most distressing item might be someone using your identity to get services, which then become billed to you. How can you prove that you didn't consume the services? I think that can be difficult, especially as we use more and more digital information that doesn't necessarily tie directly to a particular person at a point in time. This might be especially true as we store digital pictures of signatures, which often are a poor imitation of what a person's actual handwriting looks like. I shudder to think of those being used in a court of law.
Perhaps even more disconcerting is the idea that children's information is being taken at early ages, perhaps being sold decades later. I have no idea what to do here, or what I'd want others to do. These are going to be data problems we deal with for a long time, and many of us may end up collecting incorrect data in our organizations, thinking it's correct. I don't know many companies that have good processes for correcting data that's incorrect when it's received. Far too often we assume the data is correct, and only worry about ensuring the bits in a file are transformed correctly to the same bits in a database.
Security is an ongoing problem, with no easy solution. There is one thing I'm sure of. We, as data professionals, are going to be the ones frustrated by many of our efforts at security being thwarted by someone we work with.