Computational Storage Integration to Cut Host Data Transfer Load
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing capacity of storage devices leads to longer data transfer times between storage and main memory, burdening the host processor with command execution, reducing its available cycles for other tasks.
Innovation Solution
A multi-function device supports storage devices and computational storage units, with computational units hidden from the host processor, using a buffer for data sharing and asynchronous communication to reduce host involvement, and enabling near-data processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If storage device capacity is increased, then data processing capability is improved, but data transfer time to main memory increases
Solution Approach 1:
The patent introduces a computational storage unit as an intermediary component between the storage device and main memory. This unit performs data processing operations closer to the storage device, reducing the need to transfer large amounts of raw data to main memory for processing. The intermediary handles computational tasks that would otherwise require main memory access, thereby resolving the contradiction between increased storage capacity and data transfer time.
Solution Approach 2:
The system is segmented into distinct functional components: storage devices for data storage, computational storage units for data processing, and main memory for coordination. This segmentation allows each component to specialize in its function, with computational storage units handling processing operations locally rather than requiring all data to be transferred to main memory, thus reducing transfer time while maintaining high storage capacity.
2Productivity
If host processor executes commands to process data, then data processing is performed, but host processor burden increases
Solution Approach 1:
The patent extracts the data processing function from the host processor and places it in dedicated computational storage units. These units are specifically designed to handle data processing operations, freeing the host processor from this burden. The computational storage units can be configured to execute processing commands independently or with minimal host involvement, thereby maintaining high processing capability while reducing host processor complexity and load.
Solution Approach 2:
The computational storage units are designed to perform data processing operations autonomously or with minimal host processor intervention. They can execute processing commands, manage data transfers, and coordinate operations independently, enabling the system to serve itself rather than requiring continuous host processor involvement. This self-service capability maintains high productivity while significantly reducing host processor burden.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A multi-function device is disclosed. The multi-function device may include a first connector for communicating with a storage device, a second connector for communicating with a first computational storage unit, a third connector for communicating with a second computational storage unit, and a fourth connector for communicating with a host processor. The multi-function device is configured to expose the storage device and the first computational storage unit to the host processor via the fourth connector. Instead of the multi-function device comprising the second connector for communicating with the first computational storage unit, the first computational storage unit may be integrated into the multi-function device. In the latter case the multi-function device may be configured to expose the storage device and at least one of the first computational storage unit or the second computational storage unit to the host processor via the fourth connector.