Distributed Packet Processing for Network Load Balance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In current communications networks, the high load on forwarding-plane devices due to centralized context processing hampers overall network performance, as each device requires independent control module operations for modifications or upgrades, leading to complex network management.
Innovation Solution
A packet processing method and system that utilize distributed forwarding-plane devices for context processing, where ingress network elements forward packets to specific processing network elements based on flow entries containing flow description information and processing indications, allowing context processing to be performed independently across multiple devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If centralized control logic is deployed in a single control plane device, then network management efficiency is improved, but the load on forwarding-plane devices increases
Solution Approach 1:
The patent segments the centralized control logic into multiple distributed control modules deployed across different network devices. Each forwarding-plane device has its own independent control module that can be modified or upgraded separately, distributing the management load while maintaining SDN principles. This resolves the contradiction by preventing any single forwarding-plane device from bearing excessive load while still enabling efficient network management through distributed control planes.
2Device complexity
If all packets are processed by a single forwarding-plane device according to contexts, then control logic is simplified, but the forwarding-plane device becomes a bottleneck
Solution Approach 1:
The patent divides packet processing responsibilities across multiple forwarding-plane devices, each handling specific contexts or service flows. The control logic remains simplified through standardized context definitions, while processing is distributed to prevent bottlenecks. Multiple forwarding-plane devices can process different packets in parallel according to their assigned contexts, maintaining simplicity while improving throughput.
Solution Approach 2:
The patent introduces a new dimension of parallel processing by deploying multiple forwarding-plane devices that can simultaneously process different packets or context types. This transforms the single-device sequential processing model into a multi-device parallel processing architecture, eliminating bottlenecks while preserving the simplified control logic through standardized context definitions.
3Adaptability or versatility
If control modules are independently deployed in each network device, then device autonomy is improved, but network deployment and management become complex
Solution Approach 1:
The patent implements universal context definitions that can be applied across multiple forwarding-plane devices. Each device maintains autonomous control modules, but they all operate according to the same standardized context framework. This allows device autonomy while simplifying network-wide deployment, as the same context definitions can be reused across different devices rather than requiring custom configurations for each.
Data Source
AI summary
A method includes receiving, by an ingress network element, a data packet, wherein an ingress flow entry is stored in the ingress network element. The ingress flow entry comprises flow description information and a processing network element indication, where the flow description information matches the data packet, and the processing network element indication is used to indicate a processing network element that processes the data packet after the ingress network element. The method further includes sending, by the ingress network element, the data packet to the processing network element according to the ingress flow entry, so that the processing network element performs context processing on the data packet.