String Matching via Alphabet Gap Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional string matching methods in systems like intrusion detection and virus detection face inefficiencies and delays due to the need for one-to-one comparisons of large numbers of data packets and potentially harmful strings, leading to excessive system and processor resource utilization.
Innovation Solution
The system employs alphabet gap classification by separating signature strings into groups, detecting common features, and using a match module to compare input strings with these groups, allowing for efficient matching by quantifying character gaps into gap vectors and feature masks, reducing the need for character-to-character comparisons across all strings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If one-to-one comparison between signature strings and input data packets is performed, then detection accuracy is maintained, but processing time and system resources increase significantly
Solution Approach 1:
The patent segments the large set of signature strings into multiple subsets or groups. Instead of comparing each input data packet with all signature strings individually, the system divides the comparison task into smaller segments, reducing the number of comparisons needed per packet while maintaining comprehensive coverage through the segmented structure.
Solution Approach 2:
The patent performs preliminary processing on signature strings to create compressed representations or feature vectors before actual matching occurs. By pre-computing characteristics such as character frequency patterns, gap distributions, or n-gram signatures, the system reduces the complexity of real-time comparisons without sacrificing detection accuracy.
2Reliability
If one-to-one comparison between signature strings and input data packets is performed, then all potential matches are detected, but system and processor resources become unduly large
Solution Approach 1:
The patent creates compressed copies or representative vectors of signature strings that capture essential matching characteristics in a reduced form. These compressed representations serve as proxies for the full signature strings during initial filtering stages, significantly reducing memory requirements and processing overhead while preserving the ability to detect all relevant matches.
Solution Approach 2:
The patent transforms the signature matching problem from a one-dimensional character-by-character comparison into a multi-dimensional feature space. By representing signatures and data packets as vectors of characteristics (e.g., character frequencies, positional patterns, gap distributions), the system can perform efficient similarity calculations using dimensionality-reduced representations.
Data Source
AI summary
A system and method comprising providing a plurality of signature strings, inputting a plurality of strings, separating the signature strings into a plurality of signature groups, respectively detecting common features for the signature strings for each of the signature groups and the input strings, comparing the detected common features for each of the signature groups with those of the input strings, and matching the input strings and the signature groups if the detected common features for the signature groups match those of the input strings. In one aspect, the string matching further comprises matching the input strings with the signature strings for those of the signature groups having common features that match the common features of the input strings. In another aspect, the common features comprise distances or gaps between characters in the input strings and the signature strings.