Hierarchical Storage Resource Manager for Allocation Speed
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large data storage systems face scalability issues and slow resource allocation due to global resource management, leading to lengthy error recovery times and management challenges, especially when dealing with non-contiguous storage devices.
Innovation Solution
A hierarchical storage system using a resource manager and a tree-like structure to manage storage entities, allowing for efficient allocation and error recovery by scanning and summarizing metadata across levels, and dynamically managing resources by associating additional storage entities with containers as needed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If global resource pool allocation is used, then storage capacity is increased, but resource allocation speed decreases and scalability problems arise
Solution Approach 1:
The patent divides the global resource pool into hierarchical levels (global pool, volume groups, file systems, directories) where each level manages its own resources independently. This segmentation allows parallel processing and reduces the search space for allocation, improving allocation speed while maintaining total storage capacity.
Solution Approach 2:
The patent introduces a hierarchical dimension to resource management, transforming the flat global pool into a multi-level structure. This dimensional change enables resources to be allocated at the most appropriate level, reducing the complexity of global searches and improving allocation efficiency.
2Reliability
If global metadata tracking tables are maintained for consistency checking, then file system integrity is ensured, but stress on virtual memory subsystem increases and parallelization becomes difficult
Solution Approach 1:
The patent segments the global metadata tracking into hierarchical components at each level (global pool metadata, volume group metadata, file system metadata). Each level maintains its own consistency information independently, reducing the memory footprint and allowing parallel consistency checking across multiple levels.
Solution Approach 2:
The patent performs consistency checking at the most granular level necessary (individual file systems or directories) rather than requiring full global consistency checks. This partial action approach maintains file system integrity while reducing the overall complexity and memory stress.
3Productivity
If storage system is partitioned into smaller systems for faster error recovery, then fault isolation is improved, but management overhead increases especially with non-contiguous devices
Solution Approach 1:
The patent segments the storage system into hierarchical partitions (volume groups, file systems, directories) that can be independently managed and recovered. Each partition has its own metadata and resource management, enabling fault isolation and faster recovery while the hierarchical structure provides unified management across non-contiguous devices.
Solution Approach 2:
The hierarchical resource manager provides universal management capabilities across all levels of the hierarchy, handling allocation, tracking, and consistency checking uniformly whether devices are contiguous or non-contiguous. This multi-functionality reduces management overhead compared to separate management systems for each partition.
Data Source
AI summary
A system and method for storing data. In one embodiment, a storage system includes a resource manager and a hierarchical entry tree describing storage entities of the storage system. At each given level of the tree higher than the bottom level, metadata entries summarize storage availability at a level below the given level. The resource manager receives a request to store data of a target size at a target location corresponding to a first portion of the entry tree and scans the entry tree to determine if contiguous, free storage entities of the target size are available at the target location. In response to determining that contiguous, free storage entities of the target size are not available at the target location, the resource manager scans portions of the entry tree outside the first portion to identify contiguous, free storage entities of the target size, where it stores the data.