Asymmetrical Switch Connectivity Matrix for Traffic Engineering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods are inadequate for handling asymmetrical network switches, leading to potential inaccuracies and sub-optimality in path selection, inability to compute equipment disjoint paths, and inefficient routing away from failed network switches due to lack of visibility into internal device usage and varying service characteristics across different interface combinations.
Innovation Solution
An asymmetrical network switch is adapted to auto-discover and advertise a detailed connectivity matrix, including cost vectors and internal device information, allowing network path computers to select optimal paths and ensure equipment disjointness by exposing detailed connectivity attributes within the traffic engineering domain using protocols like OSPF-TE or ISIS-TE.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional connectivity matrix advertising is used for asymmetrical switches, then the network can operate with standard protocols, but path selection becomes inaccurate and suboptimal due to lack of detailed connectivity information
Solution Approach 1:
The connectivity matrix is segmented into detailed entries that represent individual internal connections within the asymmetrical switch. Each entry provides specific connectivity information for particular interface combinations, allowing path computers to make accurate routing decisions based on granular connectivity data rather than aggregated switch-level information.
Solution Approach 2:
The patent adds a new dimension of information by advertising not just which interfaces are connected, but also the specific internal devices and connection characteristics. This multi-dimensional connectivity information enables precise path computation that considers internal switch topology, resolving the information loss in conventional approaches.
2Reliability
If conventional connectivity advertising is used, then protocol simplicity is maintained, but the ability to compute equipment disjoint paths is lost
Solution Approach 1:
The connectivity matrix entries are segmented to include information about specific internal devices and connections. This segmentation allows path computers to identify and avoid shared internal resources when computing equipment disjoint paths, enabling reliability improvements without sacrificing protocol simplicity.
Solution Approach 2:
The detailed connectivity matrix acts as an intermediary that translates internal switch topology into advertisable information. It mediates between the complex internal device architecture and the path computation requirements, providing the necessary information for equipment disjoint path computation while maintaining protocol compatibility.
3Productivity
If asymmetrical switch characteristics are hidden, then network operation is simplified, but routing efficiency around failed devices is reduced
Solution Approach 1:
The detailed connectivity matrix is advertised in advance, providing path computers with complete information about internal switch connections and devices before failures occur. This preliminary information availability enables efficient real-time rerouting decisions when failures happen, improving productivity without requiring complex on-the-fly discovery mechanisms.
Solution Approach 2:
The patent changes the parameters of connectivity information from simple interface-level data to detailed entries including internal device identifiers and connection characteristics. This parameter enrichment enables efficient failure avoidance routing while the information is structured in a manageable format that doesn't excessively increase device complexity.
Data Source
Figure 1
AI summary
An asymmetrical network switch adapted (S) to auto-discover and advertise into a traffic engineering, TE, domain a switch detailed connectivity matrix, SDCM, containing for each allowed switching combination of interfaces of said asymmetrical network switch (S) at least one switch detailed connectivity matrix entry, SDCME, wherein each said SDCME represents an internal to said asymmetrical network switch potential connection interconnecting the interfaces of said interface switching combination, wherein a SDCME advertisement includes a switch detailed connectivity matrix entry cost vector, SDCME CV, which comprises a set of attributes describing cost penalties in terms of various service characteristics that a network service incurs if it selects a path or a tree traversing the asymmetrical switch (S) in accordance with the SDCME.