Multi-Level Network Packet Classification for QoS Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In enterprise network environments, it is challenging to classify and prioritize network packets effectively for Quality of Service (QoS) techniques due to varying performance requirements and characteristics such as encryption and compression, which complicates the use of existing classification methods.
Innovation Solution
A method and device for multi-level classification of network packets using packet-specific information like source and destination IP addresses, traffic direction, VLAN ID, application type, and service class, allowing deep inspection and granular classification across multiple layers of the network stack.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deep inspection and multi-level classification of network packets is performed, then classification precision and QoS optimization are improved, but device complexity and processing overhead increase
Solution Approach 1:
The classification process is divided into multiple hierarchical levels (Layer 2, Layer 3, Layer 4, and Layer 7 classification), where each level handles specific packet attributes and classification tasks. This segmentation allows the system to process different packet characteristics at appropriate layers without requiring all classification functions to operate simultaneously at every level, thereby managing complexity while maintaining high classification precision.
Solution Approach 2:
The patent introduces a multi-dimensional classification approach by examining packets across multiple OSI model layers simultaneously. Instead of single-level classification, the system evaluates packets from Layer 2 (VLAN) through Layer 7 (application layer), adding dimensional depth to the classification process. This enables more precise traffic identification and QoS policy application without simply increasing the complexity of a single classification stage.
2Measurement precision
If deep packet inspection is performed to classify encrypted or compressed traffic, then classification accuracy is improved, but processing time and loss of time increase
Solution Approach 1:
The system performs preliminary classification at lower OSI layers (Layer 2 VLAN classification, Layer 3 IP classification, Layer 4 transport classification) before attempting deeper inspection. This preliminary action establishes a hierarchical filtering mechanism where easily classifiable packets are identified and processed quickly at lower levels, reducing the number of packets that require time-consuming deep inspection at Layer 7, thereby maintaining accuracy while minimizing overall processing time.
3Reliability
If multiple classification levels and deep inspection are implemented, then traffic prioritization quality is improved, but productivity and network throughput may decrease
Solution Approach 1:
The patent applies different classification depths and inspection intensities to different types of traffic based on their specific requirements. Not all packets receive the same level of inspection - the system adapts the classification depth to the local characteristics of each packet or traffic flow. This allows critical traffic to receive detailed multi-level classification for high prioritization quality, while less critical traffic is processed more quickly at lower inspection levels, maintaining overall network throughput.
Data Source
Figure 1A
Figure 1B
Figure 1C
AI summary
The present invention is directed towards systems and methods for providing multi-level classification of a network packet. In some embodiments, network performance may be enhanced and optimized by providing QoS and acceleration engines with packet- or data-specific information. In addition to source and destination IP addresses and port numbers, packet- or data-specific information can include direction of traffic (client to host or server; server or host to client; or both), Virtual LAN (VLAN) ID, source or destination application or associated application, service class, ICA priority, type of service, differentiated service code point (DSCP), or other information. Some or all of this information may be used to classify the network packet at a plurality of layers of a network stack, allowing for deep inspection of the packet and multiple levels of granularity of classification.