Multicast Tree Reconnection via Pre-computed Backup Paths
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current media streaming technologies, such as PIM-SM and MLDP, face slow convergence and suboptimal path issues during link or node failures, leading to inefficient handling of faults in transmission trees, especially in high-bandwidth applications like HD IPTV, where fast rerouting is necessary but often results in excessive network load.
Innovation Solution
A distributed tree structure with pre-computed alternative connections allows for rapid reconnection after failures by activating dormant links, enabling data traffic to be redirected through alternate nodes, ensuring fast failure detection and full failure coverage, with nodes storing information for alternative connections and using activation packets to unblock and activate these connections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If PIM-SM or MLDP protocols are used for building multicast trees, then routing information can be obtained from unicast routing tables, but convergence time becomes slow when link or node failures occur
Solution Approach 1:
The patent pre-computes alternative upstream neighbors and backup paths before failures occur. Each router maintains a pre-calculated set of alternative upstream neighbors that can be immediately activated when a primary path fails, eliminating the need for slow reactive rerouting protocols.
Solution Approach 2:
The patent implements dynamic path switching where routers can rapidly transition between primary and alternative upstream neighbors based on real-time link status. This dynamic capability allows the multicast tree to adapt quickly to failures without waiting for protocol convergence.
2Reliability
If dual join protection technique is used to create secondary connections, then immediate alternative paths are available for fast rerouting, but network load increases significantly due to duplicate traffic
Solution Approach 1:
The patent applies protection selectively only where needed based on failure detection, rather than maintaining duplicate paths everywhere. Alternative paths are activated locally at the point of failure, allowing traffic to be routed through backup paths only when necessary, thus avoiding unnecessary network load.
Solution Approach 2:
The patent implements partial protection by maintaining alternative path information but not always activating full backup traffic flow. The system uses a threshold-based approach where alternative paths are activated only when failure is detected, avoiding the excessive action of always sending duplicate traffic.
3Stability of the object's composition
If PIM-SM waits for unicast routing recovery, then routing consistency is maintained, but real-time streaming applications suffer from service interruption
Solution Approach 1:
The patent pre-computes and caches alternative upstream neighbor information based on unicast routing tables before failures occur. This preliminary preparation allows the system to switch to alternative paths immediately when failures are detected, without waiting for unicast routing protocol convergence, thus maintaining both consistency and continuity.
Solution Approach 2:
The patent enables multicast routing to serve itself by using pre-computed alternative paths derived from unicast routing information. When failures occur, the system can independently switch to alternative paths without external intervention or waiting for unicast routing recovery, maintaining service continuity.
Data Source
Figure 1
Figure 2~3
Figure 4a
AI summary
The invention refers to operating a first node of a communication network comprising a plurality of nodes which are interconnected by communication links and wherein data traffic is provided over the communication links to provide a data stream to the nodes, wherein the first node (15) performs the steps of: detecting a data traffic interruption with respect to a second node (16), determining, if the data traffic is still received, and forwarding the data traffic to the second node (16) if the first node (15) receives the data traffic, and otherwise notifying a third node (17) of the plurality of network nodes about the data traffic interruption. The invention further refers to a corresponding node, a network and a computer program to be run in the node