Cluster Forwarding via Centralized Table Lookup
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cluster technologies in IP networks require significant software and hardware resources to maintain and search forwarding tables, which hinders their efficiency in meeting increasing network bandwidth demands.
Innovation Solution
The proposed solution involves configuring a cluster with multiple nodes where one node determines the outbound interface using a forwarding table, allowing another node to forward packets without the first node needing to maintain or search for the forwarding table, thereby reducing resource occupation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If each node in the cluster maintains and searches forwarding tables independently, then packet forwarding can be performed by any node, but software and hardware resources are significantly occupied
Solution Approach 1:
The forwarding table maintenance and search function is extracted from individual nodes and centralized in a dedicated control node. This allows forwarding decisions to be made centrally while packet forwarding itself can be distributed across multiple nodes, reducing resource occupation at each node while maintaining forwarding capability.
Solution Approach 2:
The cluster is designed with multi-functional nodes where the control node performs both control plane functions (forwarding table management) and data plane functions (packet forwarding). Other nodes primarily perform data plane functions, creating a universal system that can handle both control and forwarding tasks with optimized resource distribution.
2Reliability
If the cluster occupies more software and hardware resources, then forwarding services can be provided by all nodes, but resource efficiency decreases
Solution Approach 1:
The cluster functions are segmented into control plane tasks (forwarding table generation and updates) and data plane tasks (packet forwarding). The control plane is concentrated in one node while data plane capabilities are distributed across multiple nodes, improving resource efficiency while maintaining service availability through this functional segmentation.
Solution Approach 2:
The control node acts as an intermediary that generates and distributes forwarding tables to data plane nodes. This intermediary approach allows centralized intelligence while enabling distributed forwarding, reducing the resource burden on individual nodes while maintaining reliable forwarding services across the cluster.
Data Source
Figure 1~3
Figure 4
Figure 5~6
AI summary
Embodiments of the present invention provide a cluster that includes a first node and a second node, and the first node and the second node are configured to cooperatively perform a forwarding service on a first packet, where the first node is configured to receive the first packet by using an inbound interface and determine the inbound interface; and the second node is configured to determine an outbound interface according to a forwarding table corresponding to the forwarding service and forward the first packet by using the outbound interface of the second node. In addition, the embodiments of the present invention further provide other clusters and forwarding methods. The foregoing technical solutions help to reduce software and hardware resources occupied by a cluster.