In this podcast recorded at RSA Conference 2014, Dave Rosenberg, CTO at DB Networks, talks about a new approach to SQL injection detection and explains how they solve this problem in their products.
Signatures can’t keep pace with the attackers – either they’re too late or too broad. DB Networks technology automatically analyzes all SQL statements dispatched from your Web applications destined to your databases in order to construct a unique multi-dimensional behavioral model. Using this behavioral model, each SQL statement is then subjected to a thorough lexical analysis and SQL semantic comparison analysis. Any rogue SQL statements are immediately identified and your defined alarm procedure is invoked.
This technology is a novel patent-pending approach that learns and models an applications unique behavior for generating SQL statements. After a short learning period to construct the behavioral model, a suite of detection algorithms then evaluate each SQL statement against the application’s unique behavioral model.
New SQL statements, not seen during the learning process, go through structural analysis. Any SQL statement not consistent with the established behavioral model is identified as a likely SQL injection attack. This entire process is automatic.
DB Networks technology is completely plug-and-play — there are no blacklists, signatures, or whitelists to configure and maintain. Behavioral analysis has proven to be highly accurate at identifying even then stealthiest of SQL injection attacks.
Listen to the podcast here.