Network Search Engine Architecture With Splitting Engine And DPU Array
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional network search engines (NSEs) struggle to simultaneously achieve low power consumption, high capacity, high search throughput, and fixed search latency for all search key widths, as they typically require trade-offs between these parameters, leading to suboptimal performance in packet classification and forwarding.
Innovation Solution
The proposed NSE architecture incorporates a splitting engine and an array of data processing units (DPUs) to divide the forwarding database into bounded sub-databases, forming a hierarchical tree structure with pointer entries stored across multiple levels, allowing for efficient search operations and updates with fixed latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional network search engines use traditional routing table structures, then device complexity is reduced, but search throughput and capacity are limited
Solution Approach 1:
The routing table is divided into multiple sub-databases of bounded size, organized in a hierarchical tree structure with pointer entries. This segmentation allows parallel processing of search operations across multiple sub-databases, significantly improving search throughput while maintaining manageable complexity through structured organization
Solution Approach 2:
The patent introduces a hierarchical dimension to the traditional flat routing table structure by adding levels of pointer entries that point to sub-databases. This dimensional transformation enables more efficient search operations and better scalability without proportionally increasing device complexity
2Loss of time
If the routing table is divided into bounded sub-databases with hierarchical structure, then search latency is reduced, but device complexity increases
Solution Approach 1:
By segmenting the routing table into bounded sub-databases arranged in a hierarchical tree structure, the patent reduces search latency through parallel access to multiple sub-databases. The segmentation is designed to minimize complexity by maintaining bounded sizes and using pointer entries for efficient navigation
Solution Approach 2:
The hierarchical tree structure with pointer entries is pre-built and organized before search operations occur. This preliminary organization allows for rapid search operations by directly accessing relevant sub-databases without needing to scan the entire routing table, reducing latency while keeping the structure manageable
3Use of energy by moving object
If conventional search engines process all prefix entries, then search completeness is maintained, but power consumption increases
Solution Approach 1:
The routing table is segmented into bounded sub-databases that can be processed in parallel. This segmentation allows the system to maintain search completeness by ensuring all relevant prefix entries are covered across the hierarchical structure while reducing power consumption through efficient parallel processing and bounded processing of each sub-database
Solution Approach 2:
The patent processes only the necessary portions of the routing table (bounded sub-databases) rather than the entire table, achieving sufficient search completeness for packet forwarding while significantly reducing power consumption. The hierarchical structure ensures that partial processing of sub-databases is sufficient for correct routing decisions
4Measurement precision
If the routing table stores all network topology information, then routing accuracy is improved, but memory bandwidth requirements increase
Solution Approach 1:
The routing table is divided into bounded sub-databases organized in a hierarchical tree structure. This segmentation reduces memory bandwidth requirements by allowing selective access to relevant sub-databases based on packet destination addresses, while maintaining routing accuracy through the structured organization of topology information
Solution Approach 2:
By adding a hierarchical dimension with pointer entries, the patent enables efficient navigation and selective access to routing information. This dimensional structure reduces the amount of data that needs to be transmitted and processed, lowering memory bandwidth requirements while preserving routing accuracy through the tree-based organization
Data Source
AI summary
An improved architecture for a network search engine (NSE) is disclosed herein as including an interface manager, one or more levels of a splitting engine, an array of data processing units (DPUs), and a cascade block. A method for using the improved NSE architecture to form an efficient pointer entry database is also provided. As described herein, the improved NSE architecture simultaneously provides high speed, search throughput, update rate and capacity, coupled with low power and fixed latency searches for all search key widths.