Overlay Multicast Trees for Scalable Multi-Destination Packet Distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The inefficient distribution of multi-destination traffic in overlay networks, particularly due to the replication and forwarding of packets over tunnels, and the scalability issues arising from maintaining multicast state entries for each network device, lead to suboptimal performance and resource utilization.
Innovation Solution
Implementing a multicast tree in the underlay network with a single multicast state for each VLAN, and switching to a source-specific multicast tree based on forwarding information, reducing the need for individual tunnel replication and maintaining a scalable multicast state.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If packet replication and forwarding over tunnels is used in overlay networks, then multi-destination traffic can be distributed to multiple destinations, but the number of forwarding entries increases and network load increases
Solution Approach 1:
The patent merges multiple unicast forwarding entries into a single multicast forwarding entry. Instead of maintaining separate forwarding states for each destination in the overlay network, the system uses a single multicast group address to represent multiple destinations, thereby reducing forwarding entry complexity while maintaining adaptability for multi-destination traffic distribution.
Solution Approach 2:
The patent makes the multicast forwarding entry universal by using it to handle traffic to multiple different destinations simultaneously. A single multicast forwarding entry can serve multiple overlay destinations by mapping them to a common multicast group, providing multi-functionality that reduces the overall number of forwarding entries needed in the network device.
2Adaptability or versatility
If packet replication and forwarding over tunnels is used in overlay networks, then multi-destination traffic can be distributed, but network load increases and efficiency decreases
Solution Approach 1:
The patent performs preliminary action by establishing multicast group mappings and joining multicast groups in advance before actual traffic forwarding is needed. Network devices pre-join relevant multicast groups and set up forwarding states beforehand, so that when multi-destination traffic arrives, it can be efficiently distributed using the pre-established multicast paths rather than creating new forwarding entries on-the-fly, thereby reducing network load.
3Adaptability or versatility
If traditional overlay routing is used, then VPN over tunnels can be supported, but suboptimal path selection occurs
Solution Approach 1:
The patent introduces multicast group addresses as intermediaries between the overlay network and underlay network. Instead of directly routing packets between overlay endpoints through tunnels, the system uses multicast group addresses as intermediary destinations in the underlay network. This intermediary approach allows packets to be efficiently distributed to multiple overlay destinations simultaneously through optimized underlay multicast paths, improving speed while maintaining VPN over tunnels functionality.
Data Source
AI summary
A network device operating as a tunnel endpoint in an overlay network is provided. During operation, the network device can receive a multi-destination packet associated with a virtual local area network (VLAN) via an edge port of the network device. The network device can determine a multicast group assigned for distributing multi-destination traffic of the VLAN from a data structure in forwarding hardware of the network device. The network device can then forward the multi-destination packet via a root-path multicast tree associated with the multicast group. Subsequently, the network device can receive forwarding information shared by a second network device of the overlay network based on a control packet sent by the second network device. The network device can then program, in the forwarding hardware, a shortest-path multicast tree associated with the multicast group, wherein the shortest-path multicast tree is for distributing traffic belonging to the VLAN.