Overlay Network BGP Route Reflectors for WAN Site Connectivity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Layer 3 Virtual Private Network (VPN) services over Multi-Protocol Label Switched (MPLS) networks rely on Internet Service Providers (ISPs) for route propagation and traffic management, limiting customer control and flexibility in connecting multiple sites over a Wide Area Network (WAN).
Innovation Solution
Implementing an overlay network architecture where customer edge routers establish Border Gateway Protocol (BGP) sessions with route reflectors to announce private IP network prefixes and transport IP addresses, enabling them to learn how to reach each other and establish data paths independently, using tunneling mechanisms like GRE or L2TPv3 for secure and redundant connectivity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional MPLS-based Layer 3 VPN services are used, then guaranteed QoS and failover time are provided, but customer control and flexibility in connecting multiple sites are limited
Solution Approach 1:
The network architecture is segmented into two independent parts: the transport network (MPLS) handled by the ISP and the overlay network (customer-controlled) for site-to-site connectivity. This segmentation allows customers to independently manage their overlay network routing and connectivity while the ISP manages the underlying transport infrastructure, thus providing both flexibility and maintained reliability.
Solution Approach 2:
The patent introduces an overlay network dimension above the traditional MPLS transport layer. This additional networking dimension allows customers to implement their own routing protocols (BGP, OSPF, EIGRP) and connectivity models without being constrained by the ISP's MPLS architecture, achieving flexibility while the underlying MPLS layer continues to provide guaranteed QoS and failover.
2Adaptability or versatility
If overlay network with BGP sessions is implemented, then customer flexibility and control are enhanced, but dependency on ISP's core network remains
Solution Approach 1:
The customer network autonomously establishes BGP sessions between its own edge routers and the route reflector, announcing private IP prefixes and transport addresses without requiring ISP intervention for route propagation. The customer network self-manages its overlay routing topology, providing control and flexibility while the ISP infrastructure remains transparent.
Solution Approach 2:
A route reflector is introduced as an intermediary component that facilitates BGP route propagation between customer edge routers. The route reflector receives route announcements from edge routers, reflects them to other edge routers, and enables full mesh connectivity without requiring direct peer-to-peer configurations between all edge routers, thus reducing configuration complexity.
3Adaptability or versatility
If route reflector announces prefixes to all edge routers, then full mesh connectivity is achieved, but control plane traffic traverses ISP network
Solution Approach 1:
The control plane (BGP routing traffic) is extracted from the data plane (customer data traffic). BGP control plane traffic between edge routers and the route reflector traverses the ISP's core network for route propagation, while customer data traffic flows directly between sites over dedicated transport paths. This separation allows control plane information exchange while maintaining data plane performance and isolation.
Data Source
AI summary
A method and apparatus for connecting multiple customer sites over a wide area network (WAN) using an overlay network is described. In one embodiment of the invention, each one of multiple customer edge (CE) routers establishes a Border Gateway Protocol (BGP) session with one or more BGP route reflectors and announces their private IP network prefixes and one or more transport IP addresses to reach that CE router. The BGP route reflector(s) reflect those IP network prefixes and the one or more transport IP addresses to reach that specific CE router to the other CE routers. The CE routers receive those reflected IP network prefixes and the corresponding transport IP address(es) to reach that CE router in which those IP network prefixes belong and register them in their corresponding routing/forwarding data structures. In this way, the CE routers learn how to reach each other.