Multidimensional Network Routing via Dimensional Order and Resource Classes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing routing technologies in multidimensional networks face challenges in efficiently routing packets through congested paths and maintaining optimal latency, especially in large-scale networks with high latency channels, where traditional methods fail to dynamically adapt to changing conditions and congestion thresholds.
Innovation Solution
The method involves a routing engine that determines all possible candidate output ports for a packet in a multidimensional network, selects an optimal output port based on congestion thresholds and path weights, and routes the packet in dimensional order using two resource classes, switching between minimal and non-minimal paths as needed to avoid congestion and minimize latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If traditional routing methods are used in multidimensional networks, then the routing process is simpler, but network latency increases and congestion cannot be effectively handled
Solution Approach 1:
The routing method dynamically adapts to changing network conditions by monitoring congestion thresholds and adjusting path selection in real-time. The routing engine evaluates current network state and dynamically chooses between minimal and non-minimal paths based on congestion levels, rather than following static routing tables. This dynamic adaptation reduces latency by avoiding congested paths while maintaining manageable complexity through systematic decision rules.
Solution Approach 2:
The routing system changes the routing parameter (path selection) based on congestion threshold parameters. When congestion exceeds thresholds, the system transitions from selecting only minimal paths to incorporating non-minimal paths. This parameter-based adaptation allows the system to respond to network conditions and reduce latency without requiring complete redesign of the routing architecture.
2Speed
If minimal paths are always selected for packet routing, then the routing process is faster, but congestion and deadlocks occur in large-scale networks
Solution Approach 1:
The routing engine implements feedback by continuously monitoring network congestion levels and using this information to adjust path selection. When congestion thresholds are exceeded on minimal paths, the system receives feedback and switches to non-minimal paths. This feedback loop maintains packet routing speed by quickly responding to congestion while ensuring network stability through adaptive path diversification.
Solution Approach 2:
The system dynamically transitions between using minimal paths (for speed) and non-minimal paths (for stability) based on real-time congestion conditions. This dynamic behavior allows the network to maintain high routing speeds under normal conditions while automatically switching to more stable, less congested paths when needed, preventing deadlocks in large-scale networks.
3Reliability
If non-minimal paths are used to avoid congestion, then network reliability improves, but routing complexity and latency increase
Solution Approach 1:
The system changes the routing parameter (path length) based on congestion threshold parameters. Non-minimal paths are only selected when congestion thresholds are exceeded, not as a default. This conditional parameter change ensures reliability by avoiding congested paths while minimizing latency by sticking to minimal paths when possible, thus balancing both concerns.
Solution Approach 2:
The system applies non-minimal paths only partially - specifically when and where congestion thresholds are exceeded - rather than universally. This partial action approach maintains reliability by using alternative paths only when necessary for congestion avoidance, while minimizing the impact on routing latency by continuing to use optimal minimal paths in uncongested areas.
4Reliability
If adaptive routing with multiple resource classes is implemented, then deadlock prevention improves, but device complexity increases
Solution Approach 1:
The routing system segments the network into different resource classes (minimal paths and non-minimal paths) with distinct congestion thresholds. This segmentation allows independent control and monitoring of each path type, improving deadlock prevention by preventing circular dependencies between different path categories. The segmented approach manages complexity by creating clear, separate routing domains rather than a monolithic complex system.
Solution Approach 2:
The routing engine acts as an intermediary that mediates between minimal and non-minimal path selections based on congestion conditions. This intermediary function prevents deadlocks by controlling the interaction between different path types and ensuring proper resource class assignment. The mediator manages complexity by providing a centralized decision-making layer that simplifies the overall routing logic despite the presence of multiple resource classes.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Examples relate to routing packets in dimensional order in multidimensional networks. A packet is received at a network device in a fully connected multidimensional network and all possible candidate output ports for the packet to be routed to the destination network device with a dimensional order are determined. The candidate output ports correspond to candidate minimal paths and candidate non-minimal paths between the network device and the destination network device. An optimal output port among all the candidate output ports is selected and the packet is routed to a next hop of the network though the optimal output port using a first resource class when the optimal output port corresponds to a candidate minimal path and a second resource class when the optimal output port corresponds to a candidate non-minimal path.