SSD In-Drive Analytics Block Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current Big Data analytics approaches face bottlenecks due to the need to read large datasets from storage devices into host memory, leading to network traffic congestion and high resource requirements, making them costly and power-intensive.
Innovation Solution
Implementing in-drive analytics within solid state drives (SSDs) by dynamically switching memory blocks between I/O mode and analytics mode, allowing data analysis to occur within the SSD while maintaining concurrent I/O operations, and enabling the download of analytics logic to perform processing on unstructured data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data analytics is performed by reading large datasets from SSD into host memory, then data analysis can be conducted, but network traffic increases and host memory requirements become expensive
Solution Approach 1:
The patent extracts the data analytics function from the host system and relocates it directly to the SSD device. The SSD controller executes analytics algorithms locally on stored data without transferring data to host memory, eliminating network traffic and reducing host memory requirements while maintaining full data analysis capability
Solution Approach 2:
The SSD performs data analytics independently using its own controller and processing resources. The storage device serves its own analytics needs by executing algorithms locally, eliminating dependence on host system memory and network infrastructure for data transfer
2Productivity
If entire SSD is used for data analytics, then comprehensive analysis can be performed, but I/O operations are interrupted
Solution Approach 1:
The patent segments the SSD into multiple independent regions: analytics regions that can be dedicated to data analysis tasks and I/O regions that handle normal read/write operations. This segmentation allows comprehensive analytics to be performed on specific data portions while other regions continue to serve I/O requests without interruption
Solution Approach 2:
Different regions of the SSD are assigned different functional qualities - some regions are optimized for analytics processing while others maintain standard I/O performance characteristics. This allows the system to perform comprehensive analytics on targeted data regions without degrading the operational quality of other regions
3Productivity
If analytics algorithms are run on hosted data, then decision-making can be supported, but host system memory and processing power are required
Solution Approach 1:
The patent extracts the analytics processing capability from the host system and embeds it within the SSD controller. The SSD independently executes analytics algorithms on stored data, eliminating the need to load data into host memory or utilize host processing power while maintaining full analytics processing capability for decision-making support
Data Source
AI summary
A solid state drive (or other non-volatile storage device) includes a plurality of non-volatile storage elements arranged in blocks (or other units). Blocks (or other units) can be individually switched between analytics mode and I/O mode. The SSD (or other non-volatile storage device) performs in-drive data analytics for blocks in analytics mode while blocks in I/O mode are available for concurrent I/O operations.