Network Device Packet Classification via Field Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current packet processing methods in network devices are inefficient due to the need to compare all bits of classification information with classification rules, leading to high processing times and limited storage capacity for classification rules.
Innovation Solution
The method involves determining a target classification rule by comparing only specific bits of classification information with a reduced set of classification rules, using indication information to differentiate between match fields, thereby reducing the number of comparisons and optimizing storage efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all bits of classification information are compared with classification rules, then classification accuracy is ensured, but packet processing efficiency deteriorates
Solution Approach 1:
The patent segments the classification information into multiple fields (first field, second field, third field, etc.) and processes them in sequence. Instead of comparing all bits simultaneously, the system divides the classification task into staged comparisons across different fields, reducing the complexity of each individual comparison operation while maintaining overall classification accuracy.
Solution Approach 2:
The patent implements partial action by determining packet types using only a subset of classification information fields rather than all fields. The system performs classification based on first field, then second field, then third field in sequence, stopping when a match is found. This partial approach reduces processing time while still achieving accurate classification for most packets.
2Adaptability or versatility
If all classification rules are stored in the network device, then classification completeness is improved, but storage capacity is exhausted
Solution Approach 1:
The patent segments the classification rule set into multiple rule sets (first rule set, second rule set, third rule set, etc.), each containing a subset of classification rules. This segmentation allows the system to store and manage classification rules in smaller, more manageable units that fit within available storage capacity while maintaining comprehensive classification coverage through sequential rule set evaluation.
Solution Approach 2:
The system stores only the necessary subset of classification rules in each rule set rather than all possible rules. By organizing rules into multiple partial rule sets that are evaluated sequentially, the system achieves complete classification coverage without requiring all rules to be stored simultaneously, thus optimizing storage utilization.
3Speed
If the number of comparison operations is reduced, then packet processing speed is improved, but classification accuracy may deteriorate
Solution Approach 1:
The patent segments the classification process into multiple sequential stages, each comparing packet information against a specific rule set. The first rule set is compared first, then the second rule set, then the third rule set, and so on. This segmented approach reduces the number of comparisons needed for each stage while ensuring that if no match is found in earlier stages, subsequent stages will catch the packet, thereby maintaining overall classification accuracy.
Solution Approach 2:
The patent implements preliminary action by organizing classification rules into ordered rule sets that are evaluated in a specific sequence. Common or high-probability classification rules are placed in earlier rule sets (first rule set, second rule set) so that most packets can be classified quickly in preliminary stages. Less common rules are placed in later rule sets (third rule set, etc.), reducing the average number of comparisons needed while maintaining accuracy.
Data Source
AI summary
A packet processing method and a network device, where the method includes: receiving, by a network device, a packet, where the packet includes classification information, and where the classification information includes M fields; determining, by the network device, K fields in the M fields according to indication information stored by the network device; determining, by the network device, a target classification rule based on a first classification rule set stored by the network device and the K fields; and processing the packet according to the target classification rule.