Packet Classification Using Application Names for QoS
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In portable Internet systems, identifying the quality of service (QoS) class of IP packets and generating accurate classification rules is challenging due to limitations in conventional methods, such as reliance on protocol type and port numbers, which become dynamic and inefficient for managing varying IP addresses and traffic data.
Innovation Solution
A device and method that include a packet classification unit, QoS information storage, and a program search unit to classify packets into transport connection identifiers (CIDs) using established rules, searching for application program names in header information, and obtaining QoS parameter information to generate or change CIDs, ensuring precise classification and differentiated services.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional packet classification methods using protocol type and port number are used, then packet classification can be performed, but classification precision deteriorates because protocol type and port number are dynamic and become inaccurate for substantial traffic data
Solution Approach 1:
The patent changes the classification parameters from dynamic protocol type and port number to static application program names. By using application program names that remain constant throughout the communication session, the classification maintains high precision and reliability even for substantial traffic data. The system extracts application program names from packet headers and uses these as stable classification keys.
2Adaptability or versatility
If IP addresses and port numbers are used for classification, then packet routing can be performed, but adaptability deteriorates because IP addresses vary and require manual reconfiguration
Solution Approach 1:
The system automatically extracts application program names from packet headers and performs classification without requiring manual configuration. The access terminal autonomously identifies applications and generates classification rules, eliminating the need for manual IP address and port number configuration. This self-service approach enhances adaptability while reducing configuration complexity.
3Adaptability or versatility
If a packet classifier is connected to only one CID, then classification is simple, but versatility deteriorates because multiple QoS classes cannot be handled by a single classifier
Solution Approach 1:
The patent enables a single packet classifier to handle multiple QoS classes by connecting it to multiple CIDs. The classifier uses application program names to identify packets and routes them to appropriate CIDs based on their QoS requirements. This multi-functional approach allows one classifier to serve multiple QoS classes (UGS, rtPS, nrtPS, BE) without requiring separate classifiers for each class, thus enhancing versatility while maintaining relatively simple structure.
Data Source
AI summary
In a portable Internet system, packets are classified into a first packet classified as a transport connection identifier (CID) and a second packet that is not classified as a transport CID by using an established packet classification rule. An application program of the second packet is searched from a plurality of application programs by using the header information of the second packet in the portable Internet system. QoS parameter information for the searched application program is obtained from a plurality of QoS parameter informations corresponding to the plurality of application programs.