Split Data Plane Architecture for Flow-Based Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current flow-based switching devices face challenges in efficiently managing and routing data packets, particularly in handling unknown flows and varying traffic patterns, which can lead to increased complexity and reduced scalability.
Innovation Solution
The implementation of a split data plane architecture within flow-based switching devices, comprising a macroflow sub-plane for packet-based routing and a microflow sub-plane for flow-based routing, along with a Software Defined Networking (SDN) controller for remote management and visibility, allows for adaptable and scalable routing decisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a traditional flow-based switching device handles all routing decisions in a single plane, then routing flexibility is maintained, but device complexity and processing overhead increase significantly
Solution Approach 1:
The patent divides the data plane into two separate sub-planes: macroflow sub-plane for packet-based routing and microflow sub-plane for flow-based routing. This segmentation allows each sub-plane to handle specific routing tasks independently, reducing overall processing complexity while maintaining routing flexibility through coordinated operation of both sub-planes.
Solution Approach 2:
The patent introduces a hierarchical dimension to routing by creating multiple levels of routing planes (macroflow and microflow). This dimensional change enables the system to handle different types of traffic at appropriate granularities, reducing complexity at each level while preserving overall routing adaptability.
2Measurement precision
If a flow-based switching device uses deep packet inspection for all traffic, then routing accuracy improves, but processing time and resource consumption increase
Solution Approach 1:
The patent applies deep packet inspection selectively rather than universally. The macroflow sub-plane handles packet-based routing for traffic that doesn't require flow identification, while only the microflow sub-plane performs deep packet inspection for flow-based routing. This partial application of DPI reduces processing time while maintaining accuracy where needed.
Solution Approach 2:
Different routing processing qualities are applied to different traffic types. Packet-based routing in the macroflow sub-plane uses simpler processing, while flow-based routing in the microflow sub-plane uses deep packet inspection. This local differentiation optimizes processing time for each traffic type while maintaining necessary accuracy.
3Reliability
If a switching device maintains detailed flow state information for all flows, then routing decisions become more accurate, but memory requirements and system complexity increase
Solution Approach 1:
The patent segments flow state information storage between macroflow and microflow sub-planes. Only the microflow sub-plane maintains detailed flow state information for flows requiring flow-based routing, while the macroflow sub-plane uses simpler packet-based routing without extensive flow state tracking. This segmentation reduces overall memory requirements while maintaining routing accuracy where needed.
Data Source
AI summary
A network switching device includes a macroflow sub-plane that performs packet-based routing in the network switching device and a microflow routing module that performs flow-based routing in the network switching device. The microflow routing module is separable from the network switching device and operates to provide a data packet via a software defined network agent to an SDN controller. The data packet is associated with a first unknown flow on the network switching device. The microflow routing module further operates to receive via the SDN agent a flow-based routing rule from the SDN controller and route the data packet based upon the flow-based routing rule.