Persistency-Free Media Storage Architecture
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Maintaining up-to-date and accurate persistency databases for smooth system recovery after a crash is costly and increases software complexity, as it requires constant updates and dedicated storage for persistency data.
Innovation Solution
The method involves storing data needed for recovery as part of the main workflow, eliminating the need for persistency databases by using non-volatile storage with file tags indicating active or inactive files and associating files with interactions through headers, allowing for scanning and recovery without dedicated persistency files.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If persistency databases are maintained and constantly updated to enable smooth system recovery, then system recovery capability is improved, but software complexity and operational costs increase
Solution Approach 1:
The patent extracts the persistency database maintenance function from the system architecture, eliminating the need for dedicated persistency databases. Instead, the system uses the file system itself to store recovery information through tags and metadata embedded directly in files, thereby removing the complexity of maintaining separate persistency data structures while preserving recovery capability
Solution Approach 2:
The system enables self-service recovery by embedding recovery information directly within the files themselves through tags and metadata. When a crash occurs, the recovery mechanism scans the file system and automatically identifies active files using their embedded tags, eliminating the need for external persistency databases to track file states
2Reliability
If persistency databases are maintained and constantly updated to enable smooth system recovery, then system recovery capability is improved, but operational costs increase
Solution Approach 1:
The patent removes the operational overhead of persistency database maintenance by extracting this function entirely. The system no longer requires dedicated computing resources to update and query separate persistency databases, as recovery information is embedded directly in the file system structure through tags and metadata
Solution Approach 2:
Instead of continuously updating persistency databases during operation, the system performs recovery information updates as part of the normal file creation and modification workflow. The recovery mechanism scans files periodically or on-demand using embedded tags, eliminating continuous background processing for persistency maintenance
3Productivity
If files are scanned and tagged during routine operation to enable recovery, then recovery efficiency is improved, but file operation complexity increases
Solution Approach 1:
The patent merges the recovery information storage function with the file structure itself by embedding tags and metadata directly in files. This consolidation eliminates the need for separate persistency databases and simplifies file operations, as the file system naturally handles tag management through standard file operations
Solution Approach 2:
The file structure serves multiple functions: storing data, embedding recovery information through tags, and enabling recovery identification. This multi-functionality reduces overall system complexity by eliminating dedicated persistency storage while maintaining recovery capability
Data Source
AI summary
System and method for persistency free management of media storage including, during routine operation: continuously receiving streams of data; storing the streams of data in corresponding files in a non-volatile storage; including in the files a tag indicating whether the file is categorized as active or inactive; and when recovering from a crash: generating a list of active files by scanning the files and identifying active files. System and method for recovering after controller crash may include: during routine operation: continuously handling by the controller processes related to media metadata, by sending commands to a controlled device; sending state parameters related to the processes to the controlled device; and when recovering from the crash: retrieving the state parameters from the controlled device.