Network Path Computation via Fabric Segmentation and Stitching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional k-shortest paths algorithms struggle to keep up with the scale and complexity of modern high-scale global backbone networks, which can have billions of paths between remote devices, making real-time or near-real-time path computation difficult due to dynamic changes in network topology.
Innovation Solution
The system segments network topology into fabrics, collapses fabrics at the edge, and bundles long-haul links to perform constrained path computation in parallel, using summarized views to optimize path computation by localizing k-shortest path computation at edge nodes and caching updated paths to reduce computation frequency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If traditional k-shortest paths algorithms are used to compute paths in large-scale networks, then the number of paths found increases, but the computation time becomes too long for real-time or near-real-time operation
Solution Approach 1:
The network topology is segmented into multiple regions, with each region maintaining a simplified view of the network. Path computation is divided into two phases: first computing paths within each region using the simplified view, then stitching these regional paths together to form complete end-to-end paths. This segmentation reduces the computational complexity from considering billions of paths across the entire network to computing paths in smaller, manageable regions.
Solution Approach 2:
The system performs preliminary path computation and caching by proactively computing and storing k-shortest paths between region pairs before they are needed for actual traffic routing. These pre-computed paths are cached and can be quickly retrieved and stitched together when traffic demands arise, eliminating the need for real-time computation of all possible paths.
2Measurement precision
If the network topology is modeled in detail to ensure accurate path computation, then path accuracy improves, but the complexity of the computation system increases
Solution Approach 1:
Each region maintains a localized, simplified view of the network topology that is tailored to its specific needs and characteristics. Rather than using a single detailed global model, each region has its own customized topological view that includes only the necessary level of detail for computing paths within that region, reducing overall system complexity while maintaining local accuracy.
Solution Approach 2:
The patent introduces region boundary nodes and path stitching mechanisms as intermediaries that connect simplified regional views to the complete end-to-end path. These intermediaries enable accurate path computation across the entire network without requiring any single component to process or store the complete detailed global topology, thus reducing computation system complexity.
Data Source
AI summary
Systems and methods are disclosed for generating synthetic representations of a network and performing path computations using the synthetic representations. A model of the network is created including different representations for different regions of the network based on the network mesh patterns (e.g., a sparse mesh representation, a Clos mesh representation, and a flat mesh representation). The generated representations include synthetic, aggregated nodes and/or links in the represented region that are determined based on different processes according to the type of mesh in the region. Path computations are performed for each representation (e.g., in parallel), then joined to form end-to-end paths between a source and a destination. The computed paths may be used to select a path for routing data through the network.