Parallel Turbine TCAM FPGA Memory Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional high-performance ternary content addressable memory (TCAM) solutions are expensive and inflexible, making them unsuitable for designs with minimal cost, power, and size constraints, while FPGA implementations have higher latency and are complex to implement.

Innovation Solution

A parallel turbine TCAM design that efficiently utilizes minimal memory spaces in FPGAs for arbitrary wildcard searches, supporting variable key sizes and scalable logical lookup entries, implemented using a columnar memory structure with rotating read and write pointers for high-speed applications.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If traditional high-performance TCAM is used, then search speed and performance are improved, but cost and power consumption increase significantly

Engineering Contradiction:
Improvesearch speedVSAvoidpower consumption
Core Design Contradiction:
SpeedVSUse of energy by stationary object

Solution Approach 1:

The patent divides the TCAM into multiple banks (first bank, second bank, third bank) that can operate independently and be selectively activated. This segmentation allows the system to achieve high-speed search performance when needed while consuming less power by activating only the necessary banks, directly resolving the contradiction between search speed and power consumption.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements dynamic power management where different banks can be enabled or disabled based on operational requirements. The system can transition between high-performance mode (all banks active) and low-power mode (selective banks active), making the power consumption adaptable to actual search demands rather than continuously consuming maximum power.

Inventive Principle:
Principle #15Dynamics

2Speed

If traditional TCAM is used, then search performance is improved, but device size and cost increase

Engineering Contradiction:
Improvesearch performanceVSAvoiddevice size
Core Design Contradiction:
SpeedVSArea of stationary object

Solution Approach 1:

By dividing the TCAM into multiple smaller banks, the patent reduces the area required for each individual bank while maintaining overall search performance. The segmented structure allows for more efficient spatial utilization and reduces the total device footprint compared to a single large TCAM block.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a bank dimension to organize TCAM resources, transforming the traditional single-dimension TCAM structure into a multi-dimensional architecture. This allows searches to be distributed across multiple banks, improving performance while reducing the area required in any single dimension.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Ease of manufacture

If FPGA implementation is used, then cost is reduced, but latency increases

Engineering Contradiction:
ImprovecostVSAvoidlatency
Core Design Contradiction:
Ease of manufactureVSLoss of time

Solution Approach 1:

The patent divides the search operation into multiple parallel banks that can be activated simultaneously. This segmentation enables the FPGA to perform multiple searches in parallel, significantly reducing the effective latency compared to sequential processing, while maintaining the cost benefits of FPGA implementation.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements a continuous search mechanism where multiple banks can be searched simultaneously without idle time between operations. This eliminates gaps in the search process, maintaining continuous useful action and reducing overall latency while keeping the FPGA implementation cost-effective.

Inventive Principle:
Principle #20Continuity of useful action

4Adaptability or versatility

If traditional TCAM is used, then search flexibility is improved, but adaptability to different key sizes is reduced

Engineering Contradiction:
Improvewildcard search flexibilityVSAvoidimplementation complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent creates a universal TCAM architecture where each bank can handle different key sizes and wildcard patterns through standardized comparison logic. The same hardware structure can accommodate variable-length keys and different wildcard configurations, providing adaptability without requiring separate specialized circuits for each case.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent enables dynamic configuration of search parameters including key size and wildcard patterns by allowing the comparison logic to adapt its operation based on input requirements. This parameter flexibility allows the same hardware to efficiently handle diverse search scenarios without increasing structural complexity.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentUS9792988B2Parallel turbine ternary content addressable memory for high-speed applications
Publication Date: 2017.10.17 CIENA CORP
  • US9792988B2 patent drawing
  • US9792988B2 patent drawing
  • US9792988B2 patent drawing

AI summary

A parallel turbine ternary content addressable memory includes one or more atoms in each of one or more rows, wherein each of the one or more atoms includes a memory with N addresses and a width of M logical lookup entries, wherein N and M are integers, one or more result registers, each with a width of M, wherein a number of the one or more result registers equals a number of one or more keys each with a length of N, and a read pointer configured to cycle through a row of the N addresses per clock cycle for comparison between the M logical entries and the one or more keys with a result of the comparison stored in an associated result register for each of the one or more keys.