department of computer science and information engineering

1

Regular expression matching with input compression ： a hardware

design for use within network intrusion detection systems

Department of Computer Science and Information Engineering National Cheng Kung University, Taiwan R.O.C.

Authors: Gerald Tripp

Publisher: Journal in Computer Virology, 19 March 2007

Present: Yu-Tso Chen

Date: November, 22, 2007

2

Outline

1. Introduction 2. Definitions and problem description 3. Matching of Individual Patterns 4. Selective Grouping of Multiple

Patterns 5. Evaluation Result 6. Conclusion

3

Introduction

Build a table based automata implement-ation but to use a form of input compression

The table based approach allow the system to be dynamically updated at run time

The input compression helps to make significant reductions in the automata memory requirements.

4

Outline

1. Introduction 2. Definitions and problem

description 3. Matching of Individual Patterns 4. Selective Grouping of Multiple

Patterns 5. Evaluation Result 6. Conclusion

5

Definitions and problem description

Simple table based implementations can require quite a lot of memory resources• REs themselves that often create automata wi

th more node (and edges)

• Tables required for implementing automata can have a high level of redundancy

6

Definitions and problem description (cont.)

For a Mealy machine, the amount of memory M in bits for a DFA with s states, i input bits and o output bits

7

Outline

1. Introduction 2. Definitions and problem description 3. Regular expression

implementation 4. Input compression 5. Evaluation Result

8

Packed array DFA implementation

9

Packed transition tables

10

Packed transition tables

11

Outline

1. Introduction 2. Definitions and problem description 3. Regular expression implementation 4. Input compression 5. Evaluation Result

12

Input compression

Esn as the set of characters enabling the edge or edges between current state s and next state n

Complete set of edge sets Pa

Pa gives us the sets of characters that we are interested in for all DFA edges.• These sets may however have overlaps

13

Input compression

Pd is a set of disjoint sets of input characters

14

Example

15

Example (cont.)

16

Outline

1. Introduction 2. Definitions and problem description 3. Regular expression implementation 4. Input compression 5. Evaluation Result

17

Evaluation Result