Multicast Fast Reroute via Bidirectional Notification Packets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Protocol Independent Multicast Sparse Mode (PIM-SM) networks face slow failure reaction times due to lack of redundancy in multicast traffic routing, especially for real-time applications, and existing fast re-route mechanisms like MoFRR do not provide efficient failure detection and coverage for all failure scenarios.
Innovation Solution
The enhanced fast re-route mechanism uses downstream fast notification packets (DFNP) and upstream fast notification packets (UFNP) to reverse multicast data traffic flow between failure-detecting nodes and merge nodes, ensuring redundancy and faster failure recovery by switching to secondary paths, even for nodes without a secondary upstream multicast hop.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If PIM-SM uses a single multicast tree for distributing traffic, then the network structure is simple and easy to maintain, but there is no redundancy for re-routing multicast traffic in case of network failure
Solution Approach 1:
The patent pre-establishes a secondary multicast tree alongside the primary tree before any failure occurs. Each node along the primary path is pre-configured with backup forwarding entries pointing to the secondary tree. When a failure is detected, nodes can immediately switch to the pre-prepared secondary path without waiting for route recalculation, thus achieving fast failover while maintaining a relatively simple overall structure.
Solution Approach 2:
The patent changes the forwarding state parameters of network nodes from a single-active state to a dual-tree state with backup capabilities. By maintaining alternative forwarding entries with different next-hop parameters in the forwarding tables, nodes can dynamically switch between primary and secondary trees based on link status, providing redundancy without fundamentally altering the core PIM-SM single-tree architecture.
2Speed
If PIM-SM depends on unicast routing for multicast path construction, then the routing mechanism is simplified, but failure reaction is slow because multicast recovery must wait for unicast routing recovery
Solution Approach 1:
The patent segments the failure detection and recovery functions from the main PIM-SM control plane. By implementing independent failure detection mechanisms (such as bidirectional forwarding detection or link-state monitoring) and separate fast reroute logic in the data plane, the system can detect failures and switch paths without waiting for the slow unicast routing protocol to converge, thus accelerating failure response while keeping the core PIM-SM protocol unchanged.
Solution Approach 2:
The patent introduces an intermediary fast reroute mechanism that acts as a bridge between the primary multicast tree and secondary backup paths. This intermediary layer pre-computes alternative paths and maintains backup forwarding states, allowing nodes to quickly switch to pre-calculated secondary paths upon failure detection without directly invoking the slow unicast routing recovery process, thereby decoupling multicast failure response from unicast routing convergence.
3Reliability
If MoFRR uses a secondary path for incoming multicast stream, then immediate alternative path is provided upon primary neighbor failure, but failure detection is inefficient and not all failure scenarios are handled
Solution Approach 1:
The patent implements bidirectional feedback mechanisms where nodes send and process notification packets in both upstream and downstream directions. Downstream nodes monitor the health of upstream neighbors by expecting periodic data packets or explicit keepalive messages, and upstream nodes receive feedback about downstream status. This bidirectional feedback loop enables comprehensive failure detection along the entire multicast path, handling both upstream and downstream failure scenarios efficiently.
Solution Approach 2:
The patent makes the failure detection mechanism dynamic by implementing real-time monitoring of link and node states along the multicast path. Nodes continuously update their knowledge of neighbor status and dynamically adjust their forwarding behavior based on current network conditions. This dynamic approach allows the system to detect and respond to various failure scenarios (link failures, node failures, topology changes) in real-time rather than relying on static pre-configured failure responses.
Data Source
AI summary
An enhanced fast re-route mechanism provides increased failure coverage to a multicast communication network. If a network node detects a failure and determines that it cannot re-route multicast data, the network node sends a downstream fast notification packet (DFNP) in the network. The DFNP causes a downstream merge node to switch reception of the multicast data to its secondary path. The network node then receives an upstream fast notification packet (UFNP) from the merge node. The network node modifies its forwarding information upon receipt of the UFNP such that the multicast data is to be received by the network node from its downstream via which the UFNP was received. The DFNP and the UFNP cause the multicast data to reverse its flow direction between the network node and the merge node.