Public Review for An Improved DFA for Fast Regular Expression Matching Domenico Ficara, Stefano Giordano, Gregorio Procissi, Fabio Vitucci, Gianni Antichi, and Andrea Di Pietro Traffic classification and application identification are important functions in network administration. Traditionally, this was always done by using the port number in the transport header, but with the emerging of new applications using non standard port numbers, there is a need for another more general method. And even for legacy applications, there is a need to control the traffic to check whether it contains any harmful code as a virus or a worm. The most common approach to counter these two problems is to rely on deep packet inspection. Each undesirable (or even desirable) event is characterized by some series of characters, then the payload of packets is browsed for these specific strings seen as fingerprints of events. This method is largely used by today intrusion detection systems. The main problem with it is in the cost of browsing the memory to interpret the content of the payload and to check whether it contains some specific fingerprints or not. This is made more complex by the large number of fingerprints that exist. To
/lp/association-for-computing-machinery/an-improved-dfa-for-fast-regular-expression-matching-gqFNUnvzVs