Next Step: Data Loss Prevention
The next step up is what is generally called DLP or Data Loss Prevention, where systems monitor the network to look for data of particular sensitive types, such as credit cards or social security numbers, on the way out. I've always been leery of this sort of thing. In principle I can see how it would work, although for it to work you need many other policies in place, especially with respect to the use of encryption; the DLP needs to be able to see the data in order to analyze it. And of course many of the same companies in the encryption business are also in the DLP business. This angle on DLP also makes it a defense-in-depth component. It's not just about accidental or malicious employee actions. If an attacker were to gain access to a privileged system on the network perhaps through a targeted attack, DLP could prevent them from taking sensitive information. But the bulk of such "data leakage" cases appear to be accidental.Vendors tell me that this sort of DLP is catching on, but not as quickly. Gartner says (according to one vendor I spoke to) that 5 percent of the Fortune 500 use any real form of content monitoring. Further ratcheting up the data protection meter, we have systems designed to protect against the loss of intellectual property and other non-generic data forms, as opposed to generally well-understood forms such as credit card numbers. The appeal is obvious, but this form definitely seems nichey and I'm even more leery of it. Because of the sensitive nature of the data, you can't trust your IT people to administer the system. Various business unit personnel with authorization to access the data, or perhaps legal, or compliance officers, would need to work with the system. This alone makes things confusing. One last thing that leaves me wondering about intellectual property protection is whether we really know how effective it is. Because of the nature of the technology, it's unlikely that outsiders would be allowed to observe or audit the use of the software, and many companies would have an incentive to hush up or misrepresent results. I like the idea of DLP being mixed in with the rest of the security suite. These suites are already based around a data scanning engine, so the most efficient way to implement it is as another window on that same data scan. Better that than adding a new scanner. This whole market has the look of one taking shape over time, but you may have less time than you think, I've talked to lawyers about this stuff and their attitude is that the technology is available so you should be expected to use it. If data gets lost, or some critical IP gets stolen, you're at fault if you didn't use this newfangled tech to prevent it. Let's hope the tech picks up the pace before your own shareholders sue you. Security Center Editor Larry Seltzer has worked in and written about the computer industry since 1983. For insights on security coverage around the Web, take a look at eWEEK.com Security Center Editor Larry Seltzer's blog Cheap Hack.
I'm actually more curious about false positives than missing data. It's like a lot of security issues: A high number of false positives would be an administrative burden, either causing the alerts not to be taken seriously or for the DLP to be turned down to where it's maybe not as effective.