Partitioned Forwarding of Packets on Broadcast Links
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current layer 2 (L2) networks, such as those based on 802.1aq and TRILL, rely on a broadcast edge/routed cloud paradigm that limits packet forwarding efficiency due to the lack of forwarding adjacencies among multiple L2 devices, resulting in a single designated forwarding device handling all traffic, which can lead to suboptimal packet distribution and communication bottlenecks.
Innovation Solution
The method involves distributing the role of the designated forwarding device among multiple L2 forwarding devices by calculating a spanning tree rooted at a pseudo node shared among them, allowing each device to deterministically determine a non-overlapping partition of the forwarding space based on least cost and topology information, ensuring efficient packet forwarding without overlapping partitions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a single designated forwarding device is used to forward all packets originated on the broadcast link, then forwarding simplicity is maintained, but network capacity and forwarding efficiency deteriorate due to bottlenecks
Solution Approach 1:
The forwarding space is segmented into multiple non-overlapping partitions, with each partition assigned to a different L2 forwarding device. This segmentation allows multiple devices to share the forwarding workload simultaneously, eliminating the single-device bottleneck while maintaining deterministic packet distribution across the network
2Productivity
If multiple L2 forwarding devices participate in forwarding decisions, then forwarding efficiency improves, but determining non-overlapping partitions becomes complex
Solution Approach 1:
A pseudo node is introduced to create an equipotential reference point for all L2 forwarding devices. Each device calculates the spanning tree cost to the pseudo node, establishing a common metric that enables deterministic partition assignment without complex inter-device coordination or overlapping partition conflicts
3Ease of operation
If a broadcast edge/routed cloud paradigm is used without forwarding adjacencies, then network simplicity is maintained, but packet distribution becomes suboptimal
Solution Approach 1:
The pseudo node serves as an intermediary construct that enables multiple L2 forwarding devices to participate in forwarding decisions without requiring complex forwarding adjacencies between them. The spanning tree calculation rooted at the pseudo node provides a simple yet effective mechanism for determining packet distribution paths across the broadcast network
Data Source
AI summary
The designated forwarding device functionality for forwarding of packets originated on a broadcast link among layer 2 is shared among multiple forwarding devices of different adjacency networks. As these networks do not form adjacencies, the forwarding devices do not natively participate in a same spanning tree for determining how to forward packets, and a designated forwarding device is used for forwarding packets originated on the common broadcast link. Distributing the role of a designated forwarding device among multiple of the forwarding devices provides a means for more efficiently forwarding packets to their destinations.