Autonomous Compute Storage Device Signature Application System
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional computational storage systems face limitations due to their reliance on CPU-directed operations, requiring extensive software modifications and a limited set of developers familiar with NVMe programming. Additionally, these systems struggle with compatibility across different storage device generations and manufacturers, and require complex configuration for various deployment scenarios.
Innovation Solution
The proposed system includes a storage device chassis with a processing system and memory, which provides an autonomous compute storage device engine. This engine retrieves autonomous compute signatures and applications from a provisioning system, matches them with stored data, and executes the corresponding applications to perform compute operations, thereby offloading tasks from the CPU.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional computational storage systems use CPU-directed operations with modified software, then compute operations can be offloaded from the CPU, but the system requires extensive software modifications and a limited set of developers familiar with NVMe programming
Solution Approach 1:
The storage device autonomously executes compute operations using embedded applications and signatures stored locally, without requiring CPU-directed sequencing or extensive software modifications. The system self-services by automatically matching signatures with stored data and executing corresponding compute operations independently.
Solution Approach 2:
The patent extracts the compute operation execution logic from the CPU and embeds it directly into the storage device as autonomous applications. This separation allows the CPU to focus on high-level orchestration while the storage device handles specific compute tasks independently, reducing software complexity.
2Reliability
If conventional computational storage systems are designed for specific storage device implementations, then compute operations can be performed on specific devices, but compatibility across different storage device generations and manufacturers is limited
Solution Approach 1:
The storage device is designed with a universal architecture that can execute different compute applications across various storage device generations and manufacturers. The system uses standardized signature matching and application execution mechanisms that are not tied to specific hardware implementations, enabling broad compatibility.
Solution Approach 2:
The system dynamically adapts to different storage device implementations by loading and executing appropriate compute applications based on the specific device capabilities and stored signatures. This dynamic approach allows the same framework to operate reliably across diverse hardware platforms without requiring device-specific customization.
3Adaptability or versatility
If storage device compute applications are configured for different deployment scenarios, then specific computational storage capabilities can be provided, but complex configuration is required for various deployments
Solution Approach 1:
Compute applications and their corresponding signatures are pre-configured and stored in the storage device before deployment. This preliminary preparation allows the system to automatically execute the appropriate compute operations without requiring complex runtime configuration or manual setup for different deployment scenarios.
Solution Approach 2:
The system adapts to different deployment scenarios by changing operational parameters through the selection and execution of different compute applications rather than through complex configuration. The stored signatures and applications encode the deployment-specific parameters, simplifying the configuration process.
Data Source
AI summary
An autonomous compute storage device system includes an autonomous compute storage device signature/application provisioning system coupled to a storage device. The storage device retrieves an autonomous compute signature from the autonomous compute storage device signature/application provisioning system and, as part of a storage operation being performed in a storage subsystem in the storage device, stores data in a memory subsystem that is accessible to the storage device. If the storage device determines that the autonomous compute signature matches the data that was stored in the memory subsystem, it retrieves an autonomous compute application from the autonomous compute storage device signature/application provisioning system, and executes the autonomous compute application to perform compute operations that are associated with the data that was stored in the memory subsystem and generate at least one compute operation result.