Decoded Interval Range TCAM for Network Switch ACLs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Ternary Content Addressable Memories (TCAMs) used in Access Control Lists (ACLs) are expensive and power-intensive due to their implementation in silicon area and power consumption, limiting the number of entries and efficiency in handling wildcard bits and range matching in network switches.
Innovation Solution
The implementation of a Decoded Interval Range TCAM (DIRTCAM) that separates range search input data into groups, using Boolean functions and decoder logic to reduce the number of TCAM bits required, allowing for more efficient representation of ranges and reducing the code length by using block sizes like 2 bits or 4 bits, which halves the number of entries in the worst case.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional TCAM is used for ACL range matching, then range search capability is provided, but silicon area and power consumption increase significantly
Solution Approach 1:
The patent segments the range search problem into two separate operations: (1) exact match search for the base address using conventional TCAM, and (2) range validation using comparator logic. This segmentation allows the system to use efficient exact match hardware for address lookup while using simpler, lower-area comparator circuits for range checking, thereby reducing overall silicon area while maintaining range search capability.
Solution Approach 2:
The patent introduces an intermediary structure (address translation buffer or hash table) that maps range identifiers to actual address ranges. This intermediary allows the TCAM to store compact range identifiers rather than full range specifications, reducing the data width and associated silicon area while enabling efficient range matching through the intermediary lookup mechanism.
2Adaptability or versatility
If conventional TCAM is used for ACL range matching, then range search capability is provided, but power consumption increases significantly
Solution Approach 1:
The patent segments the power-intensive TCAM range matching operation into a two-stage process: (1) exact match lookup in the address translation buffer or hash table, and (2) comparator-based range validation. This segmentation eliminates the need for power-intensive TCAM range matching logic while preserving range search functionality, thereby significantly reducing power consumption.
Solution Approach 2:
The patent replaces the expensive, power-intensive TCAM range matching mechanism with cheaper, lower-power alternative structures such as hash tables or address translation buffers combined with comparator logic. These alternative structures consume significantly less power while providing equivalent range matching functionality.
3Measurement precision
If TCAM entries are increased to improve range representation, then range matching accuracy improves, but device complexity increases
Solution Approach 1:
The patent introduces an intermediary address translation buffer or hash table that stores the mapping between range identifiers and actual address ranges. This intermediary simplifies the overall device architecture by separating the range identification function from the range validation function, allowing the TCAM to use simple exact match entries while the intermediary handles the complexity of range management, thereby improving maintainability without sacrificing matching accuracy.
Data Source
AI summary
An embodiment of the invention includes a ternary content addressable memory (TCAM) that has input search data bits, TCAM words and range search input data bits. Each TCAM word is operable to store a match pattern and provide a match output. The match output indicates a match when the match pattern of the TCAM word matches the TCAM input search data bits. The range search input data bits are separated into groups. Each group has a bit width N where N is the number of range search input data bits. For the match pattern in each group, there is a Boolean function that uses the N range of search input data bits. (2N)/2 TCAM bits are provided for each TCAM word. 2N internal TCAM search lines are operable to search the (2N)/2 TCAM bits. Decoder logic is associated with each group that decodes the N range search input data bits.