Browse DORAS
Browse Theses
Latest Additions
Creative Commons License
Except where otherwise noted, content on this site is licensed for use under a:

High performance stride-based network payload inspection

Wang, Xiaofei (2012) High performance stride-based network payload inspection. PhD thesis, Dublin City University.

Full text available as:

PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader


There are two main drivers for network payload inspection: malicious data, attacks, virus detection in Network Intrusion Detection System (NIDS) and content detection in Data Leakage Prevention System (DLPS) or Copyright Infringement Detection System (CIDS). Network attacks are getting more and more prevalent. Traditional network firewalls can only check the packet header, but fail to detect attacks hidden in the packet payload. Therefore, the NIDS with Deep Packet Inspection (DPI) function has been developed and widely deployed. By checking each byte of a packet against the pattern set, which is called pattern matching, NIDS is able to detect the attack codes hidden in the payload. The pattern set is usually organized as a Deterministic Finite Automata (DFA). The processing time of DFA is proportional to the length of the input string, but the memory cost of a DFA is quite large. Meanwhile, the link bandwidth and the traffic of the Internet are rapidly increasing, the size of the attack signature database is also growing larger and larger due to the diversification of the attacks. Consequently, there is a strong demand for high performance and low storage cost NIDS. Traditional softwarebased and hardware-based pattern matching algorithms are have difficulty satisfying the processing speed requirement, thus high performance network payload inspection methods are needed to enable deep packet inspection at line rate. In this thesis, Stride Finite Automata (StriFA), a novel finite automata family to accelerate both string matching and regular expression matching, is presented. Compared with the conventional finite automata, which scan the entire traffic stream to locate malicious information, the StriFA only needs to scan samples of the traffic stream to find the suspicious information, thus increasing the matching speed and reducing memory requirements. Technologies such as instant messaging software (Skype, MSN) or BitTorrent file sharing methods, allow convenient sharing of information between managers, employees, customers, and partners. This, however, leads to two kinds of major security risks when exchanging data between different people: firstly, leakage of sensitive data from a company and, secondly, distribution of copyright infringing products in Peer to Peer (P2P) networks. Traditional DFA-based DPI solutions cannot be used for inspection of file distribution in P2P networks due to the potential out-of-order manner of the data delivery. To address this problem, a hybrid finite automaton called Skip-Stride-Neighbor Finite Automaton (S2NFA) is proposed to solve this problem. It combines benefits of the following three structures: 1) Skip-FA, which is used to solve the out-of-order data scanning problem; 2) Stride-DFA, which is introduced to reduce the memory usage of Skip-FA; 3) Neighbor-DFA which is based on the characteristics of Stride-DFA to get a low false positive rate at the additional cost of a small increase in memory consumption.

Item Type:Thesis (PhD)
Date of Award:November 2012
Supervisor(s):Wang, Xiaojun and Liu, Bin
Uncontrolled Keywords:Network Intrusion Detection Systems, Deep Packet Inspection, Non-deterministic Finite Automaton, Deterministic Finite Automaton, Pattern Matching, File detection
Subjects:Computer Science > Computer networks
Computer Science > Computer security
DCU Faculties and Centres:DCU Faculties and Schools > Faculty of Engineering and Computing > School of Electronic Engineering
Research Initiatives and Centres > Research Institute for Networks and Communications Engineering (RINCE)
Use License:This item is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 3.0 License. View License
ID Code:17370
Deposited On:15 Nov 2012 14:36 by Xiaojun Wang. Last Modified 15 Nov 2012 14:36

Download statistics

Archive Staff Only: edit this record