CNN-Based Data Packet Classification for Fast Rule Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data packet classification methods struggle to achieve high-speed matching and forwarding while supporting quick online updates of rules, leading to inefficiencies in systems like software-defined networking, network function virtualization, and cloud computing.
Innovation Solution
A data packet classification method based on a convolutional neural network (CNN) that merges rules by prefix combination distribution, converts rule sets into images for training, and constructs hash tables to enable efficient classification and quick rule updates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If decision tree methods are used for data packet classification, then classification speed is improved, but rule update speed deteriorates
Solution Approach 1:
The patent segments the rule set into multiple sub-rule sets based on prefix combination distribution, with each sub-rule set mapped to a dedicated hash table. This segmentation allows independent updates of individual hash tables without affecting the entire classification system, thus improving rule update speed while maintaining high classification speed through parallel hash table queries.
2Loss of time
If hash-based methods are used for data packet classification, then rule update speed is improved, but classification speed deteriorates
Solution Approach 1:
The patent introduces a new dimension by converting prefix combination distributions into image representations and using convolutional neural networks to learn optimal hash table mapping relationships. This dimensional transformation enables the system to quickly adapt hash table structures to different rule set characteristics, achieving both fast classification and fast updates by leveraging the CNN's ability to generalize from trained patterns.
3Loss of time
If T-CAM hardware is used for data packet classification, then search time is reduced, but storage space is limited and power consumption increases
Solution Approach 1:
The patent replaces the T-CAM hardware mechanism with a software-based convolutional neural network approach that runs on general-purpose processors. This substitution eliminates the storage capacity limitations and high power consumption of T-CAM while maintaining fast search performance through the CNN's learned hash table mapping, which can dynamically adapt to any rule set size without hardware constraints.
4Device complexity
If dimensionality reduction methods are used for data packet classification, then rule complexity is reduced, but rule update speed deteriorates due to multiple rule table updates
Solution Approach 1:
The patent extracts the prefix combination distribution characteristics from the rule set and uses them as the sole basis for generating hash table mappings, eliminating the need to update multiple dimension-specific rule tables. By taking out and focusing on this key characteristic, the system reduces rule complexity while enabling fast updates through single-hash-table modifications.
Data Source
AI summary
The disclosure provides a data packet classification method and system based on a convolutional neural network including merging each rule set in a training rule set to form a plurality of merging schemes, and determining an optimal merging scheme for each rule set in the training rule set on the basis of performance evaluation; converting a prefix combination distribution of each rule set in the training rule set and a target rule set into an image, and training a convolutional neural network model by taking the image and the corresponding optimal merging scheme as features; and classifying the target rule set on the basis of image similarity, and constructing a corresponding hash table for data packet classification.