Compressed Binary Tree Search for Faster Traffic Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing traffic classification methods using binary decision trees face issues with deep tree structures, excessive memory usage due to range rule conversions, and rule replication, which affect decision efficiency and search speed.
Innovation Solution
The method involves determining compression parameters to reduce tree depth, segmenting rules using bit selection algorithms, converting range rules to prefixes when efficient, and extracting replicated rules to create separate sub-rule sets, thereby compressing binary trees and optimizing search processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If range rules are converted to prefixes during binary tree creation, then the tree structure can be formed, but memory space is excessively occupied due to rule expansion
Solution Approach 1:
The patent segments the binary tree structure into multiple layers with different compression ratios. Different types of nodes (compression nodes vs. regular nodes) are segmented and handled differently. Compression nodes store only compression parameters and bitmaps, while regular nodes store complete rule information, thereby reducing overall memory occupation while maintaining tree creation capability
Solution Approach 2:
The patent changes the parameter representation in different tree nodes. Instead of storing complete prefix rules in all nodes, compression nodes use compressed parameters (compression ratio, bitmap) to represent multiple rules. This parameter transformation reduces memory usage from storing full rule sets to storing compact compression metadata
2Measurement precision
If the binary tree is created with detailed rule segmentation, then classification accuracy is maintained, but tree depth becomes large affecting decision efficiency
Solution Approach 1:
The patent implements a nested structure where compression nodes contain multiple child nodes that collectively represent compressed rules. This nesting allows the tree to maintain detailed classification paths (for accuracy) while reducing the number of visible levels (for efficiency). The compression node acts as a container that expands only when necessary during traversal
Solution Approach 2:
The patent adds a compression dimension to the traditional binary tree structure. Instead of only depth and width dimensions, compression nodes introduce a compression ratio dimension. This allows the tree to achieve better classification accuracy by maintaining detailed paths where needed while reducing tree depth through compression, thereby improving decision efficiency
3Loss of substance
If multiple binary decision trees are created for different sub-rule sets, then rule replication is reduced, but device complexity increases
Solution Approach 1:
The patent creates a universal compressed binary tree structure that can handle multiple sub-rule sets within a single tree. The compression nodes and bitmaps provide a multi-functional framework that can represent different rule sets without requiring separate trees for each sub-rule set, thereby reducing device complexity while maintaining the benefit of reduced rule replication
Data Source
AI summary
Methods and devices for creating, compressing and searching a binary tree are provided by the present disclosure, which relates to the field of traffic classification. Binary tree compression includes: determining a number of a compression layer or a number of an intermediate node, compressing a binary tree according to the number of the compression layer or the number of the intermediate node to form at least one compression node, and creating a bitmap of the compression node. Therefore, the depth of the decision tree is reduced, the search speed is increased, rule expansion is avoided, and rule replication is reduced.