Upon receiving malware detection rules that are to be identified with respect to an input traffic stream, a rule database that requires less storage capacity than the malware detection rules is generated by substituting tokens for selected symbol strings within the malware detection rules. A compressed traffic stream is generated by substituting the tokens for instances of the selected symbol strings within the input traffic stream, and then compared with the rule database to determine whether the input traffic stream contains one or more symbol sequences that correspond to any of the malware detection rules.