State Machine Engine Automata Processor Graph Traversal
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional computers face inefficiencies in performing complex pattern recognition tasks due to the increasing volume of data and number of patterns to be identified, leading to bottlenecks in processing time, especially when searching for spam or malware in data streams.
Innovation Solution
A state machine engine with a hierarchical configuration of finite state machine lattices operates in parallel, analyzing data streams across multiple criteria simultaneously, mimicking the biological brain's hierarchical neuron layers to enhance processing speed and efficiency in pattern recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional computers search for patterns one at a time in a data stream, then the system can process each pattern thoroughly, but the processing time increases and creates a computing bottleneck
Solution Approach 1:
The patent divides the pattern recognition task into multiple independent circuit elements, each responsible for detecting a specific pattern or portion of a pattern. These circuits operate in parallel to process different patterns simultaneously, eliminating the sequential bottleneck while maintaining detection accuracy through dedicated circuitry for each pattern type.
Solution Approach 2:
The patent transitions from sequential single-dimensional pattern processing to parallel multi-dimensional processing by deploying multiple circuit elements across different spatial dimensions. This allows simultaneous detection of multiple patterns across the data stream, effectively adding a parallel processing dimension that resolves the time-delay contradiction.
2Productivity
If hardware distributes the data stream among a plurality of circuits to search for patterns in parallel, then processing speed increases, but the intermediate results become larger than the original input data
Solution Approach 1:
The patent merges the results from multiple parallel circuit elements through logical operations that combine detections efficiently. Rather than maintaining separate large result sets from each circuit, the system consolidates intermediate results by identifying and combining matching patterns, reducing the overall volume of intermediate data while preserving all necessary detection information.
Solution Approach 2:
The circuit elements are designed with multi-functional capabilities to handle multiple pattern types and detection scenarios. This universality allows the same circuit architecture to process various patterns simultaneously, reducing the need for dedicated specialized circuits for each pattern type and thereby minimizing intermediate data volume while maintaining high processing speed.
3Adaptability or versatility
If the number of patterns to be identified increases with the variety of spam and malware, then the system can detect more types of threats, but the delay before the system is ready to search the next portion of the data stream increases
Solution Approach 1:
The patent implements preliminary action by pre-configuring multiple circuit elements to recognize different pattern types simultaneously. Rather than preparing to search for each new pattern type sequentially, the system has multiple circuits already prepared and operational, each capable of detecting specific pattern varieties. This allows the system to handle increased pattern variety without increasing preparation delay, as all detection capabilities are simultaneously available.
Data Source
AI summary
An apparatus includes a state machine engine. The state machine engine may also include an automaton, whereby the automaton is configured to analyze data from a query related to solving a graph. The automaton may further be configured to report an event representative of a satisfaction of a node solving a hop of the graph by a portion of the input data stream.