VXLAN Tunnel Fabric for Scalable Layer-2 Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Network bridges face scalability, resiliency, and efficiency issues due to extensive flooding, single link connectivity, and inability to utilize multiple shortest paths, which limits their ability to support high-sectional traffic and results in half-utilized network segments.
Innovation Solution
A virtual-extensible-local-area-network (VXLAN) tunnel is established between nodes in different layer-two networks, using Ethernet over IP/GRE tunnels and a modified IS-IS protocol to propagate link state database information and discover network topology, enabling multiple shortest paths without loops and forming a virtual bridging domain that appears as a single network device.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If network bridges use extensive flooding to locate unknown devices, then device location can be found, but network scalability deteriorates
Solution Approach 1:
The patent introduces a control plane intermediary that separates the data plane (forwarding) from the control plane (decision-making). The control plane uses IS-IS routing protocol to distribute link-state database information and calculate shortest paths centrally, eliminating the need for extensive flooding at the data plane level. This intermediary control mechanism enables scalable network operation while maintaining device location capabilities.
Solution Approach 2:
The patent changes the operational parameters of bridge-like devices by transitioning from traditional MAC address-based flooding to IP address-based shortest path routing. By using IS-IS protocol and link-state database information, the system calculates optimal paths based on network topology parameters rather than relying on extensive flooding, thereby improving scalability while maintaining location capabilities.
2Device complexity
If network bridges use single link connectivity, then device simplicity is maintained, but network resiliency deteriorates
Solution Approach 1:
The patent implements dynamic path selection by enabling bridge-like devices to calculate and switch between multiple shortest paths based on real-time network topology information from the link-state database. This dynamic capability allows the network to adapt to link failures and maintain resiliency without significantly increasing device complexity, as the path calculation is performed using standardized IS-IS protocol mechanisms.
Solution Approach 2:
The control plane performs preliminary actions by pre-calculating shortest paths and distributing link-state database information before failures occur. This advance preparation enables rapid failover and path switching when link failures happen, improving network resiliency without requiring complex real-time decision-making at the data plane level.
3Stability of the object's composition
If network bridges force single link connectivity, then topology disturbance is reduced, but ability to use multiple shortest paths deteriorates
Solution Approach 1:
The patent segments the network functionality by separating control plane operations (IS-IS protocol, link-state database, shortest path calculation) from data plane operations (packet forwarding). This segmentation allows the data plane to maintain stable, simple forwarding behavior while the control plane dynamically determines multiple shortest paths, thereby achieving both topology stability and improved traffic utilization through efficient path selection.
Data Source
AI summary
In some embodiments, a method includes defining, by a processor included in a first node, a virtual-extensible-local-area-network (VXLAN) tunnel between the first node included in a first layer-two network, and a second node included in a second layer-two network, the VXLAN tunnel traversing at least one node of a layer-three network. The method includes receiving, at the first node, a layer-two data unit that is sent from a third node included in the first layer-two network, to a fourth node included in the second layer-two network. The method includes encapsulating, at the first node, the layer-two data unit to define an encapsulated data unit that includes a VXLAN header. The method includes sending the encapsulated packet from the first node towards the fourth node via the VXLAN tunnel.