NVMe SSD Controller Failover via Bitmap Redirection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current multipath storage systems do not provide an alternative method for accessing Persistent Memory Regions (PMR)/Controller Memory Buffer (CMB) regions during failover situations, limiting data access across host systems when one host system experiences physical access failure with Non-Volatile Memory Express Solid State Drive (NVMe-SSD).
Innovation Solution
A method and apparatus that allow a first host system to access the PMR/CMB region of a second host system by sending vendor-specific commands to modify a bitmap mapping unique controller identifiers and base address registers, enabling redirection of read/write packets to access the memory region of the second host system's SSD controller.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a host system experiences physical access failure with NVMe-SSD, then data access is lost, but existing multipath systems cannot provide alternative access to memory regions across different host systems
Solution Approach 1:
The patent enables SSD controllers to perform multiple functions: serving their native host system and alternatively serving other host systems' memory regions during failover. The bitmap data structure and vendor-specific commands allow the controller to dynamically redirect read/write packets to access memory regions belonging to different hosts, making the storage path universal across multiple hosts.
Solution Approach 2:
The patent introduces vendor-specific commands and a bitmap data structure as intermediaries between host systems and SSD controllers. These intermediaries enable the controller to understand and process cross-host memory region access requests, translating host failover needs into controller-specific redirection operations without requiring changes to the underlying NVMe protocol.
2Adaptability or versatility
If namespace access is enabled across multiple hosts, then data accessibility improves, but memory region access across hosts remains unsupported
Solution Approach 1:
The patent segments the storage address space into identifiable memory regions associated with specific hosts using a bitmap data structure. Each bit in the bitmap corresponds to a specific host's memory region, allowing the SSD controller to selectively redirect access to the appropriate region based on the failing host's identity, enabling precise cross-host memory region access.
Solution Approach 2:
The patent performs preliminary configuration by establishing bitmap mappings between host identifiers and memory region base addresses before failover occurs. This pre-configuration allows the SSD controller to immediately redirect traffic upon detecting a host failure without requiring real-time discovery or negotiation, ensuring rapid failover to alternative memory regions.
3Reliability
If traditional multipath failover is implemented, then path redundancy is achieved, but memory region access across different hosts is not possible
Solution Approach 1:
The patent implements dynamic failover by enabling the SSD controller to adaptively redirect access requests based on real-time host availability. The bitmap data structure and vendor-specific commands allow the controller to dynamically update its routing decisions, switching between serving its native host and alternative hosts' memory regions as conditions change, making the system flexible rather than static.
Data Source
AI summary
A first host system in a multipath storage system acts on behalf of a second host system when the second host system is unable to access a solid state drive (SSD). The first host system configures a bitmap table using an SSD controller of the first host system to access memory of the SSD associated with the second host system. The memory accessed on behalf of the second host system may be in a region including persistent memory, base address register memory and/or controller memory buffer of the second host system.