SIEM Ain't DAM

I've been getting questions about the difference between system information and event management (SIEM) and database activity monitoring (DAM) platforms. It's easy to get confused given their similarities in architecture. There's also a great deal of overlap in events that each collects and the way they handle information. Couple that with aggressive marketing claims, and it seems impossible to differentiate between the two platforms.If you look at the architectural diagrams for SIEM and DAM, you will see both are designed for central event aggregation, normalization, analysis, and reporting. Both collect database audit logs and, in at least one case, SQL activity on the database network port. Finally, both offer compliance solutions and rules, and report out of the box to "accelerate compliance," "aid security operations," and "reduce costs."

Other than these commonalities, SIEM and DAM are completely different. The difference is the degree of analysis and the focus of the analysis.

DAM provides far more in-depth analysis on database traffic. It performs extensive profiling of individual database queries as well as query traffic in general. DAM uses content, context, and behavioral analysis techniques. With more database-specific information, coupled with more analysis techniques, much finer grained policies can be implemented. And DAM performs most analysis in real-time.

SIEM vendors advertise they have database monitoring, but only a couple have viable monitoring solutions. Most provide auditing, but fall down on monitoring databases. (That doesn't mean DAM is a better technology -- it's just the degree of focus.)

Where SIEM platforms shine is in their ability to examine all devices and applications, and to correlate more events from wider variety of sources. Their ability to collect a picture of events across the entire IT infrastructure is unparalleled. Finally, the largest DAM installations process only a fraction of the events a normal SIEM or log management system will. DAM is not geared for the demands of log management in the size and scale SIEM platforms must provide day in and day out.

SIEM and log management platforms are far more advanced in normalization, correlation, and managing large data sets. They have the capability to provide forensic snapshots far outside the scope of DAM products. They are commonly capable of accepting data, logs, and events from every single device and application in the IT environment. SIEM systems are really focused on large-scale aggregation and correlation of generic events, while DAM is focused on real-time analysis of database transaction. For general forensic analysis, SIEM is the product of choice.

Some of you might think the differences in these two well-known solutions are straightforward, with little point in contrasting, but run through the common business problems and you'll see the difficulty. If you need to meet PCI-DSS requirements, do you select the product that tells you your network is secure, or the one that tells your credit cards are secure? Do you select the product that tracks and monitors access to network resources, or the one that tracks and monitors access to cardholder data? Which tool best associates user identity to machines? User identity to applications? Which is best at restricting access to card holder data? Which is best at identifying vulnerabilities? Dig in; the answers could surprise you.

If you are looking at these technologies and trying to figure out which is best for your security challenge, then focus on your use cases. Don't assume DAM is the best option just because you need to review database audit logs. Similarly, SIEM provides a good overview of events, but don't assume it's going to keep your database secure. You need to dig into the goals and use cases before you can identify the correct technology. Or you might find you need both.

Adrian Lane is an analyst/CTO with Securosis LLC, an independent security consulting practice. Special to Dark Reading.