Network Switch Flow Filter Using Incomplete Comparison Table
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional network switch technologies require significant memory space to store data flow records and incur additional processing loads due to the need to analyze and classify all incoming data flows, leading to inefficiencies in identifying network traffic types and prioritizing services.
Innovation Solution
Implementing a packet filtering mechanism with an incomplete comparison table, such as a Bloom filter, to reduce memory usage and processing loads by quickly verifying data flows and only processing those that do not match existing flow entries or filtering conditions, thereby directing new flows to a flow analyzing module for classification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a network switch stores all data flow records in a flow table for complete comparison, then identification accuracy is improved, but memory space requirement increases significantly
Solution Approach 1:
The patent divides the flow table into two parts: a flow filter (incomplete comparison table) and a flow table (complete comparison table). The flow filter stores only essential flow information with a bit array, while the flow table stores complete flow records. This segmentation reduces memory space requirement while maintaining identification accuracy for known flows.
Solution Approach 2:
The patent extracts the comparison function from the traditional flow table and implements it through a flow filter using a bit array. The flow filter extracts only the necessary flow identification capability, separating it from the complete flow record storage, thereby reducing memory usage while maintaining identification precision.
2Measurement precision
If the network switch analyzes and classifies all incoming data flows, then traffic identification capability is improved, but processing load increases
Solution Approach 1:
The patent performs preliminary action by using the flow filter to pre-check incoming data flows against stored flow information before routing. This preliminary comparison filters out known flows that don't require analysis, reducing the processing load on the flow analyzing module while maintaining identification capability.
Solution Approach 2:
The patent extracts the flow analysis function from all incoming packets and applies it selectively only to packets that fail the flow filter check. This extraction reduces processing load by avoiding unnecessary analysis of known flows while maintaining traffic identification capability for unknown flows.
3Measurement precision
If the network switch uses traditional flow table for complete comparison, then classification accuracy is improved, but device complexity increases
Solution Approach 1:
The patent segments the flow table into a flow filter (incomplete comparison table) and a flow table (complete comparison table). The flow filter handles simple bit array comparisons for known flows, while the flow table handles complete flow records. This segmentation reduces device complexity by using simpler data structures for common cases.
Solution Approach 2:
The patent uses a simple bit array structure for the flow filter, which is a simpler and more memory-efficient alternative to the traditional flow table. This cheap structure handles the common case of known flow identification, reducing overall device complexity while maintaining classification accuracy.
Data Source
AI summary
A method and a system for processing a data flow with an incomplete comparison process are provided. The method is implemented by a network device that includes a flow table and a flow filter in a memory thereof. A flow analyzing module is provided for analyzing and classifying packets of an input flow, and identifying an application category to which the input flow belongs. The flow table is queried according to a result of resolving the input flow for determining whether the input flow matches any flow entry of the flow table. The flow filter is queried if the input flow fails to match any flow entry of the flow table for determining whether features of the input flow match conditions of the flow filter. The input flow is processed accordingly, without needing to copy all flows that do not match the flow entries to the flow table.