Adaptive Data Request Interleaving for Storage Interface Stalls
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In data storage systems, request buffers can become filled when all requests are directed to the same block of host data, leading to delays and inefficiencies, as other requests cannot be sent until at least one request from that block is completed, which increases cost and complexity by enlarging the request queue.
Innovation Solution
Implementing an interleaving scheme where requests for portions of data from multiple blocks of host data are sent in an interleaved order, allowing other requests to be completed even if one block is delayed, with the interleaving scheme chosen based on metrics such as the number of unfulfilled requests, block capacity, or host memory characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the request queue is enlarged to hold more requests, then the risk of interface stalls is reduced, but the cost and complexity of the system increases
Solution Approach 1:
The patent segments the request queue into multiple independent sub-queues, each associated with a specific host memory block. This segmentation allows requests for different blocks to be processed independently, preventing a single block's delays from stalling the entire interface while maintaining a compact overall queue structure.
Solution Approach 2:
The patent implements dynamic selection of interleaving schemes based on current system conditions such as the number of unfulfilled requests and block capacity. This dynamic adaptation allows the system to optimize performance without requiring a permanently oversized request queue, balancing reliability and complexity adaptively.
2Productivity
If all requests are directed to the same block of host data, then the request processing is simplified, but the interface efficiency deteriorates due to stalls
Solution Approach 1:
The patent segments requests by their target host memory blocks, organizing them into separate sub-queues. This allows the system to process requests from multiple blocks concurrently through interleaved access patterns, preventing interface stalls caused by concentrating all requests on a single slow block while maintaining organized request management.
Solution Approach 2:
The patent changes the parameter of request ordering from sequential (all requests to one block completed before moving to the next) to interleaved (requests from multiple blocks alternated). This parameter change transforms the system from being vulnerable to single-block delays to efficiently utilizing multiple blocks in parallel, improving overall interface productivity.
Data Source
AI summary
An example of a system includes a host interface, a set of non-volatile memory cells, and one or more control circuits coupled to the host interface and coupled to the set of non-volatile memory cells. The one or more control circuits are configured to access a host memory through the host interface by sending host memory access requests for two or more blocks of host data according to an interleaving scheme.


