摘要 |
The occurrence of false positives and the post-processing of digital streams subjected to examination by a deterministic finite state machine for character strings are reduced by combining location-based pattern matching, e.g. on packet headers, and content-based pattern matching, e.g. on payloads of packets. One scheme allows automatic transition from a header match state into an initial state of a content matching machine. Another scheme is based on a rules graph defining strings of match states and the examination of a list of match states (rather than characters) which have been previously determined, for example by means of header matching and content matching. The latter is also capable of comparing offset and depth values associated with the match states with offset and depth criteria.
|