Computational Storage Message Routing for CPU Bottleneck Relief
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional network-attached storage devices inefficiently process data messages through a central processing unit (CPU), leading to bottlenecks in processing power and communication bandwidth, which limits the scalability of storage capacity.
Innovation Solution
Implementing separate processing paths for control and data messages, where control messages are routed through a processing device and data messages are directly communicated between a storage client and storage device without CPU intervention, reducing the CPU's workload.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data messages are processed through a central processing unit (CPU), then administrative control and security are maintained, but processing power bottlenecks and communication bandwidth limitations occur
Solution Approach 1:
The patent segments message processing into two distinct paths: control messages are routed through the CPU for administrative processing, while data messages are routed directly between the storage client and storage device without CPU intervention. This segmentation resolves the contradiction by allowing the CPU to maintain security and control functions while eliminating it as a bottleneck for data processing operations.
2Ease of operation
If data messages are processed through a central processing unit (CPU), then centralized control is maintained, but communication bandwidth requirements increase
Solution Approach 1:
The patent extracts data message processing from the CPU's processing path, creating a direct communication channel between the storage client and storage device. This extraction maintains centralized control through the control message path while eliminating the bandwidth consumption associated with routing data messages through the CPU.
3Adaptability or versatility
If separate processing paths are implemented for control and data messages, then scalability of storage capacity is improved, but device complexity increases
Solution Approach 1:
The patent implements segmentation of message processing paths with distinct routing logic for control and data messages. This segmentation enables scalability by allowing independent optimization of each path while the added complexity is managed through structured routing rules and message classification mechanisms.
Data Source
AI summary
A computing device having: a local host system running a data application; and a storage product. The storage product has: a local storage device; a network interface to receive storage access messages from a remote host system; and a processing device configured to identify, from storage access messages, a first subset for processing by the data application, and a second subset bypassing the local host system. The first subset of the storage access messages includes first write messages configured by the remote host system to write first data into the storage product. The data application running in the local host system can generate second data based at least in part on the first data. The storage product can write the second data into the local storage device in response to the first write messages.