Transactional Memory Storage Allocation via Linked Lists
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing memory allocation algorithms are inefficient for managing file and storage space, leading to fragmentation, poor performance, and unsuitability for managing storage devices due to their reliance on memory-oriented linked lists, which are slow for file access.
Innovation Solution
A system and method for transactional memory and storage allocation that uses storage-oriented linked lists, aligns memory and storage space in large chunks, and minimizes overhead by storing management structures within a single page, allowing for efficient allocation, resizing, and deallocation of memory and storage space simultaneously.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If memory allocation algorithms are used for storage management, then ease of operation is improved, but speed deteriorates due to reliance on memory-oriented linked lists
Solution Approach 1:
The patent changes the fundamental parameter of data structure organization from memory-oriented linked lists to storage-oriented linked lists. This parameter change enables the system to maintain the ease of operation provided by linked list management while achieving storage device speeds by organizing data structures to match storage device characteristics rather than memory characteristics.
Solution Approach 2:
The patent inverts the conventional approach by using storage-oriented linked lists instead of memory-oriented linked lists for storage management. This inversion allows the system to apply familiar allocation algorithms in a reversed context, achieving both ease of operation and storage device speed compatibility.
2Speed
If storage-oriented linked lists are used, then speed is improved, but device complexity increases
Solution Approach 1:
The patent creates a universal allocation system that handles both memory and storage through a common interface. The storage-oriented linked lists serve multiple functions: they enable fast storage access, maintain allocation tracking, and provide a unified interface for both memory and storage management, thereby reducing overall system complexity despite the specialized data structure.
Solution Approach 2:
The patent segments the storage management system into distinct components: storage-oriented linked lists for fast access, allocation tracking structures for memory management, and a unified allocation interface. This segmentation allows each component to be optimized independently while working together to reduce overall device complexity.
3Productivity
If memory and storage space are aligned in large chunks, then productivity is improved, but loss of time increases due to larger allocation units
Solution Approach 1:
The patent segments large storage chunks into smaller allocatable units through the storage-oriented linked list structure. While large chunks are allocated for efficiency, the linked list enables fine-grained management within those chunks, allowing the system to recover and reallocate small spaces efficiently, thereby reducing the time loss from large allocation units.
Solution Approach 2:
The patent implements dynamic allocation within large storage chunks using storage-oriented linked lists. This allows the system to initially allocate large chunks for productivity but dynamically manage and reallocate smaller units within those chunks as needed, reducing the time loss associated with large fixed allocation units.
Data Source
AI summary
An integrated system for transactionally managing main memory and storage devices derived from the interfaces and methodologies historically associated with dynamic memory allocation. The methodology has a wide range of applicability including areas such as hardware storage devices (i.e. firmware), operating system internals (i.e. file systems) and end-user software systems.