Recognition of Computer Viruses by Detecting Their Gene of Self Replication
Abstract
An approach to the detection of malicious software by detecting its ability to self-replicate is proposed, implemented and tested. The approach is justified by the following realities most malicious programs propagate themselves through the Internet to maximize the impact of the information attack; self-replication of legitimate programs is quite uncommon; number of practical self-replication techniques is quite limited and is to be repeatedly used by new malicious codes. A Source Code Analyzer operating as a specialized compiler (interpreter) and a special syntax library were developed for the detection of self-replication functionality in source codes /scripts prior to execution. Major building blocks of the existing self-replication techniques were defined in the domain of system calls and their attributes, and a procedure for the reconstruction of these blocks by analyzing the flow of system call was established. A dynamic Code Analyzer and System Calls Monitor were developed for the run-time detection of the attempted self-replication in executable and encrypted executable codes. The efficiency of the developed technology, including the ability to detect previously unknown malicious programs has been experimentally demonstrated.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 2006
- Accession Number
- ADA448622
Entities
People
- Douglas Summerville
- Victor Skormin
Organizations
- Binghamton University