Multicast Tree Link Utilization Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network flow scheduling methods, such as shortest path flow scheduling, often lead to congested network spots and reduced throughput due to inefficient link utilization, which can block new flow establishment and increase operational costs.
Innovation Solution
A method for network flow scheduling that establishes a multicast tree by minimizing link utilization and tree depth, adjusting links based on utilization objectives and constraints, and re-evaluating the tree cost to optimize link usage and avoid congested links, using techniques like mixed integer linear programming and software-defined networking.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If shortest path flow scheduling is used, then flow establishment speed is improved, but link utilization deteriorates causing congestion
Solution Approach 1:
The patent changes the routing parameters from purely shortest-path-based to a multi-parameter optimization approach that considers link utilization, tree depth, and cost metrics. This allows the system to select paths that balance speed of establishment with long-term network health, avoiding congestion on heavily used links.
Solution Approach 2:
The patent implements dynamic multicast tree adjustment where the routing decisions are not static but adapt based on current network conditions. The system continuously evaluates link utilization and adjusts the multicast tree structure to redistribute traffic away from congested links while maintaining efficient flow establishment.
2Quantity of substance
If multicast tree minimizes number of links, then link cost is reduced, but link utilization increases causing bottlenecks
Solution Approach 1:
The patent transforms the optimization problem from minimizing a single parameter (number of links) to a multi-parameter optimization that balances link count, link utilization, and tree depth. This allows the system to accept slightly more links when necessary to avoid creating utilization bottlenecks.
Solution Approach 2:
The patent applies different optimization criteria to different parts of the multicast tree. Rather than uniformly minimizing all links, the system identifies and protects specific links from becoming bottlenecks by adjusting the tree structure locally to distribute traffic more evenly across available paths.
3Length of moving object
If multicast tree depth is minimized, then routing efficiency is improved, but link capacity constraints are violated
Solution Approach 1:
The patent changes the optimization approach from minimizing tree depth in isolation to a constrained optimization where tree depth is minimized subject to link capacity constraints. This ensures that the resulting multicast tree both maintains routing efficiency and respects the physical limitations of network links.
4Device complexity
If conventional flow scheduling is used, then implementation simplicity is maintained, but network throughput is reduced
Solution Approach 1:
The patent introduces a multicast tree optimization layer that acts as an intermediary between simple flow scheduling requests and the complex network routing infrastructure. This intermediary handles the computational complexity of multi-parameter optimization while presenting a simplified interface for flow establishment, thereby improving throughput without requiring complete redesign of the scheduling system.
Data Source
AI summary
Embodiments are provided for path flow scheduling of multicast traffic through a network. The paths for traffic flow are determined to optimize link utilization in terms of bandwidth and link capacity, and limit link cost. In an embodiment, a method is implemented for network flow scheduling. The method includes establishing, by a controller of a network, a multicast tree which includes a plurality of links for sending multicast traffic from a source to multiple destinations. The tree is established based on minimizing a number of links in the multicast tree. The tree is then adjusted by replacing one or more of the plurality of links to reduce the link utilization. The tree adjustment is repeated by further replacing one or more links in the multicast tree to further reduce the link utilization.