Multicast Path Determination via Spanning Tree
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing solutions for determining end-to-end paths in networks are inefficient, requiring numerous data packets and relying on packet entropy and hashing, which consume significant network bandwidth and computational resources.
Innovation Solution
The method involves sending a single data packet from a sending node to a multicast IP address, where intermediary nodes determine and send copies of the packet along multiple downstream paths, allowing all possible end-to-end paths to be determined with minimal packet transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional path determination methods are used, then all end-to-end paths can be determined, but the number of data packets required increases to n^2, consuming significant network bandwidth and computational resources
Solution Approach 1:
The patent merges multiple path determination operations into a single data packet by implementing a spanning tree protocol where one root node sends one packet that traverses all possible paths through the network. Intermediate nodes replicate and forward the packet along multiple downstream paths, combining the functionality of n^2 separate packets into one unified measurement operation.
Solution Approach 2:
The single data packet serves multiple functions simultaneously: it performs path discovery, network topology mapping, and reachability testing across all n nodes. The packet structure includes fields for tracking visited nodes and maintaining spanning tree relationships, allowing one packet to accomplish what traditionally required n^2 separate probing operations.
2Loss of information
If multiple data packets are sent to determine all paths, then comprehensive path information is obtained, but network bandwidth and computational resources are significantly consumed
Solution Approach 1:
The patent combines multiple information-gathering operations into a single data packet transmission. The spanning tree protocol ensures that one packet sent from the root node traverses all n-1 other nodes through intermediate replications, collecting complete path information while using only 1/n the bandwidth of traditional methods.
Solution Approach 2:
Intermediate nodes create and forward copies of the single incoming packet along multiple downstream paths. This copying mechanism allows the original information to be distributed throughout the network without requiring multiple original transmissions, efficiently capturing all path information through systematic replication of the single probe packet.
3Reliability
If traditional path determination is used, then path availability is monitored, but the system requires a controller device and numerous packets, reducing scalability
Solution Approach 1:
The spanning tree protocol implements self-service path determination where each intermediate node autonomously processes the incoming packet, determines its own position in the spanning tree, and forwards copies to appropriate downstream nodes. This distributed self-organization eliminates the need for a centralized controller device, allowing the network to automatically discover and monitor all paths through decentralized node behavior.
Solution Approach 2:
The protocol establishes a spanning tree structure as a preliminary framework before actual path determination occurs. This pre-established hierarchical structure organizes node relationships and defines forwarding paths, enabling efficient single-packet traversal and eliminating the need for complex real-time controller intervention during path discovery operations.
Data Source
AI summary
A sending node may send a network message to a multicast IP address. The sending node and a recipient node may be in communication via a plurality of intermediary nodes. The recipient node may receive a number of copies of the network message that corresponds to a number of equal cost paths between the sending node and the recipient node. The recipient node may send a reply to each of the copies of the network message to the sending node. The sending node may provide the replies to a collector module, which may use the replies for network analytics, network monitoring, machine learning, and the like.