Maximally Disjoint Trees for Packet Switched Network Load Balancing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current communication systems in packet switched networks face challenges in efficiently balancing load across multiple paths, leading to congestion and suboptimal network utilization, particularly due to the limitations of equal cost multi-path routing which is topology-dependent and may not utilize all available paths.
Innovation Solution
The implementation of maximally disjoint trees in communication systems, where each tree is mutually exclusive from others, allows for load balancing by computing a set of these trees based on network topology and using them to distribute packets across paths, ensuring fair distribution and optimal resource utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If equal cost multi-path routing is used, then load balancing is achieved, but network topology dependency limits path utilization
Solution Approach 1:
The network topology is segmented into multiple maximally disjoint trees, where each tree represents a separate set of mutually exclusive paths from source to destination nodes. This segmentation allows the system to utilize all available paths in the network rather than being constrained to equal cost paths only, thereby resolving the contradiction between throughput and path selection flexibility.
Solution Approach 2:
The invention changes the routing parameter from cost-based selection (ECMP) to tree-based disjoint path selection. By computing maximally disjoint trees based on network topology, the system can select paths that are mutually exclusive and maximize resource utilization, overcoming the limitations of topology-dependent equal cost routing.
2Productivity
If multiple paths are utilized for load balancing, then network utilization improves, but path disjointness may not be maximized
Solution Approach 1:
The system performs preliminary computation of maximally disjoint trees based on network topology before actual packet forwarding. By pre-computing the set of mutually exclusive paths and storing them in routing tables, the system maximizes resource utilization without adding real-time complexity to packet forwarding operations.
3Productivity
If maximally disjoint trees are computed, then all available paths are utilized, but computation complexity increases
Solution Approach 1:
The routing system automatically computes maximally disjoint trees using the existing network topology information that is already available in the network. By leveraging the topology data that routers already maintain for routing purposes, the system achieves comprehensive path utilization without requiring additional complex topology discovery mechanisms.
Data Source
AI summary
Various example embodiments for supporting load balancing in packet switched networks are presented herein. Various example embodiments for supporting load balancing in packet switched networks may be configured to support load balancing in packet switched networks based on use of disjoint trees. Various example embodiments for supporting load balancing in packet switched networks may be configured to support load balancing in packet switched networks based on use of maximally disjoint trees. Various example embodiments for supporting load balancing in packet switched networks based on use of maximally disjoint trees may be configured to support load balancing in packet switched networks using per-flow load balancing, per-packet load balancing, randomized load balancing (RLB), or the like, as well as various combinations thereof.