TCAM Entry Coding for On-the-Fly Error Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Ternary Content Addressable Memories (TCAMs) face challenges in error detection and correction due to their ternary nature, making it difficult to implement reliable error correction techniques without excessive overhead or restrictive usage limitations, especially in sensitive networking applications prone to single-event upsets.
Innovation Solution
The implementation of error detection and correction codes that encode values in a way to ensure a minimum distance between encoded states, allowing for single-error correction and double-error detection, along with segregated entry groups and independent encoding for each entry group, enables reliable operation and tolerance for errors in TCAMs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional error detection and correction techniques are used in TCAMs, then reliability is improved, but device complexity and overhead increase significantly
Solution Approach 1:
The patent divides the TCAM into multiple banks, with each bank containing a subset of the encoded comparison values. This segmentation allows error correction to be performed independently in each bank, reducing the overall complexity and overhead compared to encoding the entire TCAM content in a single error correction code structure.
2Reliability
If error correction codes are implemented in TCAMs, then reliability is improved, but restrictions are imposed on where values can be stored
Solution Approach 1:
The patent employs a universal error correction code structure that can be applied to any comparison value stored in any bank of the TCAM. The encoding scheme and error correction logic are designed to work uniformly across all banks, allowing values to be stored and corrected in any location without requiring location-specific error correction mechanisms.
3Reliability
If background scrubbing is used to detect errors in TCAMs, then some errors are detected, but errors may not be detected before errored data is used
Solution Approach 1:
The patent encodes the comparison values with error detection and correction codes before they are stored in the TCAM banks. This preliminary encoding ensures that when values are read or searched, the error correction can be performed immediately on-the-fly, detecting and correcting errors before the data is used for routing decisions, eliminating the delay associated with background scrubbing operations.
Data Source
AI summary
A Content Addressable Memory (CAM) or Ternary CAM (TCAM) provides error detection and correction (EDAC). EDAC codes are chosen based on logical and physical properties of the CAM/TCAM. An entry in the CAM/TCAM comprises a plurality of groups, each group comprising a plurality of storage bits. Writes to the storage bits are encoded to enable EDAC. Lookup data is divided into lookup groups of one or more bits, and is applied to corresponding groups of entries to be searched. In one embodiment, storage bits in a group are first decoded to detect and/or to correct errors and then compared with a lookup group to produce a hit indication. In another embodiment, storage bits in a group are logically combined with a lookup group to produce a hit indication, wherein a correctable error in the storage bits does not affect correctness of the hit indication.