Regional Route Reflectors for Multi-Hop SD-WAN Gateways
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing routing solutions for large-scale SD-WAN deployments face limitations such as end-to-end visibility loss during handoffs, constrained maximum hops, and manual controller assignment, which hinder scalability and connectivity across geographical regions.
Innovation Solution
Implementing a distributed, disjoint gateway router model with multi-hop routing support, branch-to-branch VPN, and customizable VPN profiles, along with seamless switching between redundant transit points using route summarization, to enable full-mesh or customizable mesh redundancy and resiliency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a common controller model is used for route exchanges, then route propagation is simplified, but the maximum supported overlay hops is limited to two, constraining hierarchical deployments
Solution Approach 1:
The patent segments the common controller into multiple regional controllers (route reflectors), each responsible for a specific geographic region. This segmentation allows hierarchical deployments with multiple overlay hops while maintaining simplified route propagation within each region. Edge routers connect to their regional controller, and regional controllers peer with each other, enabling multi-hop support without complex end-to-end controller management.
Solution Approach 2:
The patent introduces a hierarchical dimension to the controller architecture, organizing controllers into regional units that can be stacked vertically. This dimensional change from a flat single-controller model to a hierarchical multi-controller model enables support for more overlay hops while maintaining operational simplicity at each level through standardized regional interfaces.
2Reliability
If manual assignment of controllers to edges is performed, then the number of edges connecting to a controller stays within acceptable limits, but deployment complexity increases
Solution Approach 1:
The patent implements self-service through automated controller assignment where edge routers autonomously select their regional controller based on geographic location or network topology. The system automatically distributes edges across multiple controllers without manual intervention, using load-balancing mechanisms and automated provisioning to maintain connection limits while simplifying deployment procedures.
3Speed
If underlays are used for routing, then direct connectivity is achieved, but end-to-end visibility is lost when overlay to underlay handoffs occur
Solution Approach 1:
The patent introduces route reflectors as intermediary controllers that maintain visibility of all edge routers in the network. When overlay-to-underlay handoffs occur, the route reflector acts as a mediator that tracks and maintains end-to-end routing information, ensuring visibility is preserved while still enabling fast underlay routing through the overlay control plane.
4Productivity
If distributed gateway routers are deployed, then scalability is improved for large-scale deployments, but system complexity increases
Solution Approach 1:
The patent segments the gateway function into distributed regional controllers that operate independently within their regions. This segmentation enables linear scalability where new regions and edge routers can be added without reconfiguring the entire system. Each regional controller maintains a manageable subset of routing information, reducing individual device complexity while increasing overall system capacity.
Solution Approach 2:
The patent applies homogeneity by standardizing the regional controller architecture across all regions, using identical software images and configuration templates. This homogeneous design simplifies deployment and operations despite the distributed nature of the system, as all controllers behave predictably and can be managed through standardized procedures.
Data Source
AI summary
Some embodiments of the invention provide a method for implementing an SD-WAN connecting multiple sites at multiple physical locations. The method is performed at a first route reflector for a first region of the SD-WAN. The method receives, from a hub router of the first region, a peer-connection notification regarding a newly connected first edge router located at a first site in a second region. The method determines that a routing table maintained by the first route reflector does not include routes of the first edge router and that the first route reflector does not have a direct connection to the first edge router. Based on said determining, the method requests routes of the first edge router from the hub router. After receiving from the hub router the requested routes of the first edge router, the method updates the routing table to include the routes of the first edge router.