Active-Membership I/O Routing for Synchronous Storage Replication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage systems face inefficiencies in data management and reliability, particularly in flash storage systems, due to unnecessary write operations and lack of centralized control over storage processes, leading to reduced reliability and increased wear on storage devices.
Innovation Solution
Implementing a direct-mapped flash storage system where the operating system directly addresses data blocks without translation by storage controllers, offloading device management responsibilities, and utilizing non-volatile RAM for quick data buffering and power failure protection, along with dual storage array controllers for failover and distributed management of erase blocks and metadata.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If storage controllers translate and manage data blocks, then device management is centralized, but write operations become redundant and storage device wear increases
Solution Approach 1:
The patent extracts the data block translation and management functions from storage controllers and relocates them to the operating system. This eliminates redundant write operations where data is written both to cache and to storage devices, as the OS directly manages data block placement. The storage controllers are reduced to simple I/O path components, no longer performing translation operations that caused the contradiction.
Solution Approach 2:
The operating system takes on the responsibility of directly addressing and managing data blocks on storage devices, making the system self-sufficient in data management. This eliminates the need for storage controllers to perform translation operations, allowing the system to manage its own data blocks efficiently without intermediary translation layers that caused redundant writes.
2Extent of automation
If storage controllers manage data blocks, then data management is automated, but data integrity and reliability decrease
Solution Approach 1:
The patent removes data block management responsibilities from storage controllers and transfers them to the operating system. This extraction allows the OS to implement more reliable data management protocols and directly verify data integrity before writes, eliminating the automated but flawed translation processes that previously compromised data integrity.
3Speed
If data is buffered in volatile memory, then quick data access is achieved, but power failure results in data loss
Solution Approach 1:
The patent implements preliminary action by flushing buffered data to non-volatile storage before power failure can occur. The system proactively writes data from volatile cache memory to non-volatile storage devices, ensuring data is safely persisted before any potential power loss event, thus resolving the contradiction between fast access and power failure protection.
Solution Approach 2:
The patent provides beforehand cushioning by maintaining data copies in non-volatile storage before power failure occurs. This creates a safety buffer where data is protected in advance, cushioning the system against the harmful effect of power failure and preventing data loss that would otherwise occur with volatile memory alone.
Data Source
AI summary
Determining active membership among a set of storage systems, including: determining, by a cloud-based storage system among the set of storage systems, that a membership event corresponds to a change in membership to the set of storage systems synchronously replicating the dataset; applying, in dependence upon the membership event, one or more membership protocols to determine a new set of storage systems to synchronously replicate the dataset; and for one or more I/O operations directed to the dataset, applying the one or more I/O operations to the dataset synchronously replicated by the new set of storage systems.