Storage Device Offloading Data Processing to Reduce Host Workload
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In storage systems, the workload of host devices becomes severe due to managing and processing data across multiple storage devices, leading to inefficiencies in data handling and communication.
Innovation Solution
A storage device and system architecture where the storage devices perform offloaded work such as data processing, compression, and parity generation independently, reducing the load on the host device by using a first interface to receive and process data chunks, and a second storage device to generate parity for data recovery, with peer-to-peer communication between storage devices.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the host device manages and processes data across multiple storage devices, then data management capability is improved, but the workload of the host device becomes severe
Solution Approach 1:
The patent extracts data processing tasks from the host device and relocates them to storage devices. The storage device performs data processing, compression, and parity generation independently, separating the host's management function from the processing function to reduce host workload.
Solution Approach 2:
The storage device performs self-processing by generating processed data and parity information internally without requiring continuous host intervention. The operation circuit processes data chunks and generates parity automatically, enabling the storage system to serve itself.
2Adaptability or versatility
If the host device processes and manages data to be stored in multiple storage devices, then data management capability is improved, but the workload of the host device becomes severe
Solution Approach 1:
The patent extracts complex data processing operations from the host device and implements them within the storage device. The operation circuit handles data processing, compression, and parity generation locally, reducing the complexity burden on the host device.
3Productivity
If storage devices communicate with the host device for data processing, then data handling efficiency is improved, but the amount of data exchanged with the host increases
Solution Approach 1:
The patent extracts data processing operations from the host-device communication loop and implements them within the storage device itself. The operation circuit processes data locally and only exchanges necessary control signals with the host, significantly reducing the quantity of data transferred.
Solution Approach 2:
The storage device performs self-processing of data including compression and parity generation without requiring continuous host involvement. The first interface outputs processed data and signals to the host only when necessary, enabling autonomous operation and reducing communication overhead.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Described is a storage device that includes a first interface, an operation circuit, and a nonvolatile memory. The first interface receives a first data chunk from a host device. The operation circuit generates first processed data by processing the first data chunk and generates a first signal indicating a size of the first processed data. The nonvolatile memory stores the first processed data in a storage location, when the storage location at which the first processed data are to be stored is designated to the storage device based on the first signal. The first interface outputs the first signal to the host device.