Statistical-based intrusion detection
On January 24, 2003, the W32.SQLExp.Worm (later named Slammer/Sapphire) was released into the wild. This worm exploited a stack-based buffer overflow vulnerability in Microsoft's SQL Server 2000 software (including MSDE 2000). While vulnerabilities affecting Microsoft products are nothing new, the speed at which this worm propagated was extremely novel - scary in fact. The worm was released and within ten minutes it had compromised 90% of all vulnerable systems worldwide. Before this incident, worms of this type were merely theoretical, given serious consideration primarily in the academia.
It takes even the fastest vendors hours or days to produce a signature for rule-based intrusion detection (RBID) systems. In the case of this worm however, a vulnerable network would be compromised in a matter of seconds, much too quickly for even the most diligently updated RBID system. So what is the solution to a worm that doubles its infection rate every 8.5 seconds? The answer may lie in a lesser-known intrusion detection method called statistical-based (also referred to as behavior-based) anomaly detection.
[ Read more ]
Reading our newsletter every Monday will keep you up-to-date with security news.
Receive a daily digest of the latest security news.