Web applications are constantly exposed to security threats from code injection attacks such as SQL injection in cyberspace.At present

most detection methods against SQL injection attacks have low execution efficiency and low detection accuracy

and are not easy to be reused.According to the characteristics of injection vulnerability

a SQL injection behavior detection method based on aspect-oriented programming and dynamic taint analysis is proposed

the taint analysis process is encapsulated by the aspect unit

so that the typical program crosscutting-concerns are separated from the base system

which improves the reusability of detection code.The Advice mechanism is used to dynamically load the various detection component implementations to execute the detection code at runtime to counter typical code injection attacks such as SQL injection against Web applications.Experiments show that this method can realize the self-protection process without modifying the application execution engine and source code

so as to effectively defend against seven typical types of SQL injection attacks such as tautologies

logically incorrect queries

union query

piggy-backed queries

stored procedures

inference query

alternate encodings

and so on.