Granular Network Traffic Tracking via FEC Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network vendors face challenges in scaling their solutions to accommodate higher data rates and denser network interfaces for tracking non-sampled network traffic, leading to substantial costs due to the large number of transistors and memories required.
Innovation Solution
A method and system for programming network elements to track network traffic by receiving an accounting policy configuration specifying network prefixes, storing forwarding equivalence class (FEC) and forwarding information base (FIB) entries, enabling granular tracking and segmentation of network traffic, thereby allowing for more precise accounting statistics and telemetry.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If large numbers of transistors and memories are deployed to track non-sampled network traffic, then tracking capability is improved, but device complexity and cost increase substantially
Solution Approach 1:
The patent segments network traffic tracking by dividing the destination address space into multiple network prefixes. Each prefix is associated with a separate FEC (Forwarding Equivalence Class) entry in the FIB table, allowing granular tracking of traffic to different destinations without requiring comprehensive hardware resources for all possible traffic types simultaneously.
Solution Approach 2:
The patent introduces a new dimension of organization by mapping network prefixes to FEC indices, which then map to BNHI (Bridging Next Hop Information). This multi-level mapping structure allows efficient traffic classification and tracking using existing hardware resources arranged in a hierarchical dimension rather than requiring flat comprehensive hardware coverage.
2Measurement precision
If large numbers of transistors and memories are deployed to track non-sampled network traffic, then accounting precision is improved, but cost increases substantially
Solution Approach 1:
The patent segments accounting tracking by network prefixes, allowing precise accounting statistics for specific destination ranges without requiring expensive comprehensive hardware resources. The FIB table structure enables selective tracking of traffic to configured prefixes, providing precise accounting where needed while avoiding unnecessary costs for tracking all traffic types.
Solution Approach 2:
The patent changes the parameter of traffic classification from binary (sampled vs non-sampled) to multi-level classification based on network prefixes. This allows precise accounting statistics to be generated for specific prefix ranges using software-based FEC table lookups rather than expensive dedicated hardware counters for all traffic types.
3Ease of manufacture
If existing hardware resources are used for traffic tracking, then cost is reduced, but scalability to higher data rates is limited
Solution Approach 1:
The patent implements dynamic scalability through the FIB table structure that can be configured and updated based on network needs. The mapping between network prefixes, FEC indices, and BNHI can be dynamically adjusted to accommodate varying data rates and traffic patterns without requiring proportional increases in hardware resources, enabling cost-effective scalability.
Solution Approach 2:
The patent organizes hardware resources in a hierarchical dimensional structure (network prefixes → FEC indices → BNHI) that scales efficiently with data rates. This multi-level mapping allows the system to handle higher data rates by leveraging the existing hardware architecture in a dimensionally organized manner rather than requiring linear scaling of resources.
Data Source
AI summary
In general, embodiments of the invention relate to a method for programming a network element to granularly track network traffic traversing through one or more network elements.