Network Switch Interconnected Member Nodes Forwarding Path Calculation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network switch architectures require each node to perform separate forwarding path calculations for packet forwarding, leading to increased computational processing and power consumption, which can hinder efficient packet switching.
Innovation Solution
Implementing a front plane stacking architecture where interconnected physical switches function as a single virtual switch, with a topology manager in each node that detects network topology, calculates forwarding paths, and programs ASICs to minimize redundant processing, allowing packets to be forwarded efficiently without recalculating paths at each node.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If each node performs separate forwarding path calculations for packet forwarding, then forwarding decisions can be made independently at each node, but computational processing and power consumption increase
Solution Approach 1:
Multiple physical switches are merged into a single virtual switch through front plane stacking, where a central controller consolidates the forwarding path calculation functionality. Instead of each node performing independent calculations, the controller computes forwarding paths centrally and distributes them to member nodes, reducing redundant computational processing and power consumption while maintaining coordinated forwarding decisions across the stacked system.
2Extent of automation
If each node performs separate forwarding path calculations, then forwarding can be decentralized, but computational processing increases
Solution Approach 1:
A central controller acts as an intermediary between the external network and member nodes in the stacked switch. The controller receives packets, performs centralized forwarding path calculations, and directs member nodes to forward packets according to pre-computed paths. This intermediary approach maintains automated forwarding capability while eliminating redundant computational processing at each node, as member nodes simply execute instructions from the controller rather than performing independent calculations.
3Adaptability or versatility
If forwarding paths are calculated at each node, then routing flexibility is maintained, but processing time increases
Solution Approach 1:
The central controller performs preliminary forwarding path calculations and distributes routing information to all member nodes in advance. When packets arrive at member nodes, they can be forwarded immediately using pre-computed forwarding tables without requiring real-time path calculation. This preliminary action maintains routing flexibility through centralized intelligence while dramatically reducing processing time at individual nodes, as the computational burden is shifted to advance preparation by the controller.
Data Source
AI summary
According to examples, a network switch may include plurality of member nodes interconnected by member links to exchange data plane traffic. The network switch may also include a commander node to run control and management plane protocols for the plurality of member nodes. A first member node of the plurality of member nodes that receives a packet incoming into the network switch calculates a forwarding path for the packet to a destination through a second member node of the plurality of member nodes and forwards the packet to the second member node, and the second member node forwards the packet according to the forwarding path calculated by the first member node without calculating another forwarding path for the packet.