User Plane Function Packet Processing via Hash Table Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current communication networks face inefficiencies in processing large volumes of data due to excessive signaling between control and user planes, particularly in 5G core networks, leading to capacity limitations and bottlenecks in managing user sessions and packet detection rules, which result in suboptimal throughput and increased resource consumption.
Innovation Solution
Implementing programmable network devices with embedded hardware such as FPGAs and GPUs that perform rapid lookups of packet data rules using hash tables, allowing for targeted processing of user-plane data packets based on precedence, thereby reducing the need to evaluate all packet detection rules and minimizing system resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all packet detection rules are evaluated for each session, then packet processing accuracy is improved, but processing speed deteriorates
Solution Approach 1:
The patent segments the packet detection rules into session-specific rules stored in hash tables, allowing the system to evaluate only relevant rules for each packet rather than all rules. This segmentation enables rapid lookup by hashing packet characteristics to identify the appropriate session and its associated rules, thereby maintaining accuracy while improving speed.
Solution Approach 2:
The patent performs preliminary actions by pre-storing session information and packet detection rules in hash tables during session establishment. This preliminary organization of data allows for rapid retrieval and evaluation during packet processing, eliminating the need to evaluate all rules from scratch for each packet.
2Adaptability or versatility
If the number of packet filters for packet detection is increased, then packet detection capability is improved, but system resource consumption increases
Solution Approach 1:
The patent segments packet detection rules into session-specific groups stored in hash tables, allowing the system to maintain a large number of packet filters without proportionally increasing resource consumption. Only the rules relevant to active sessions are loaded and evaluated, reducing memory usage and processing overhead while maintaining comprehensive detection capability.
Solution Approach 2:
The patent changes the organizational parameter of packet detection rules from a flat structure to a hash table structure keyed on session identifiers. This parameter change enables efficient retrieval and selective evaluation of rules, reducing the effective number of rules processed at any given time while maintaining the ability to handle a large total number of filters.
3Productivity
If control plane and user plane are separated, then network scalability is improved, but signaling overhead increases
Solution Approach 1:
The patent performs preliminary actions during session establishment by pre-configuring session-specific packet detection rules in the user plane function. This allows the user plane to autonomously process packets without requiring continuous control plane intervention, reducing signaling overhead while maintaining scalability.
Solution Approach 2:
The patent enables the user plane function to self-service by storing and evaluating session-specific rules locally in hash tables. This eliminates the need for the control plane to evaluate every packet, allowing the user plane to independently make forwarding decisions and reducing control plane signaling overhead.
Data Source
AI summary
Programmable networking devices configured to perform various packet processing functions for packet filtration, control and user plane separation (CUPS), user plane function (UPF), pipeline processing, etc. Upon arrival of a user plane packet, a UPF performs a rapid lookup or hash table of the provisioned PDRs associated with a given PFCP session, arrange PDRs in decreasing order of precedence, and process the packet more efficiently than evaluating all PDRs.