Co-located SGW PGW PDR Linking for Throughput Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The combined Serving Gateway (SGW)/Packet Gateway (PGW) architecture experiences inefficiencies due to overlapping functionalities and unnecessary hops in data flow, leading to suboptimal throughput in current systems.
Innovation Solution
The method involves linking Packet Detection Rules (PDRs) for co-located SGW/PGW systems, optimizing data-path by performing Packet Forwarding Control Protocol (PFCP) sessions based on Tunnel Identifier End Point (TEID) and User Equipment (UE) IP, and using PDRs to determine local next hops, thereby eliminating unnecessary S5-U GTPU header operations and checksum calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional SGW/PGW architecture with separate entities is used, then network functionality and reliability are maintained, but data transmission throughput is reduced due to unnecessary hops and redundant operations
Solution Approach 1:
The patent combines SGW and PGW into a single co-located entity, merging their functionalities to eliminate unnecessary hops between separate SGW and PGW nodes. This consolidation directly improves data transmission throughput by reducing latency and redundant processing while maintaining the distinct functional roles of SGW (mobility management) and PGW (packet routing) within the same physical infrastructure.
2Productivity
If PDR linking is implemented in co-located SGW/PGW, then throughput is optimized by eliminating redundant operations, but processing complexity at each node increases
Solution Approach 1:
The patent implements preliminary action by pre-establishing PDR linking relationships between SGW and PGW before data transmission occurs. The PDRs are configured in advance to identify co-located SGW/PGW pairs and establish direct forwarding paths, eliminating the need for dynamic route discovery and redundant header operations during actual data transmission, thus optimizing throughput while managing processing complexity.
3Productivity
If unnecessary S5-U GTPU header operations are performed, then protocol compatibility is maintained, but transmission efficiency is reduced due to redundant encapsulation and decapsulation
Solution Approach 1:
The patent extracts and eliminates unnecessary S5-U GTPU header operations from the data path when SGW and PGW are co-located. By identifying the co-location scenario, the system removes redundant encapsulation and decapsulation steps that would otherwise be required for inter-node communication, directly improving transmission efficiency while maintaining protocol compatibility through conditional application of optimization.
4Reliability
If checksum calculations are performed at each hop, then data integrity is verified, but processing time increases due to redundant calculations
Solution Approach 1:
The patent merges checksum calculation operations by performing them once at the source node rather than redundantly at each intermediate hop. When SGW and PGW are co-located, the system identifies that intermediate checksum calculations are unnecessary since the data remains within the same physical infrastructure, thereby reducing processing time while maintaining data integrity through single-point verification.
Data Source
AI summary
A method for linking Packet Detection Rules (PDRs) for optimizing throughput of combined Serving Gateway (SGW)/Packet Gateway (PGW) architecture is disclosed. In one embodiment the method includes, determining a SGW and a PGW are co-located on a system and when the SGW and PGW are co-located on a same system then performing a Packet Forwarding Control Protocol (PFCP) session lookup for an uplink packet at the SGW; determining which PDR to use for sending the uplink traffic; using the PDR to remove any tunneling protocol header; checking the PDR to determine if the next hop is local; when the next hop is local, then finding the PGW PDRs for this FTEID; identifying a PGW PDR matching with the current packet; performing any packet encapsulation; and after performing any optional header checksum calculations, sending the packet.


