Aligned Variable Reclaim Process for EFI NVRAM
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current EFI implementations require additional NVRAM for fault-tolerant variable and data reclamation, leading to increased costs and performance degradation, especially in complex server systems where real-time responses are critical.
Innovation Solution
The Aligned Variable Reclaim Process (AVRP) technique uses system management memory (SMRAM) to shadow NVRAM, ensuring memory-aligned transactions and data borrowing to reduce the time required for fault-tolerant variable reclamation without additional NVRAM, thereby improving performance and reducing costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If additional NVRAM is used for fault-tolerant variable reclamation, then reliability is improved, but cost increases and performance degrades
Solution Approach 1:
The patent introduces an intermediary mechanism using system management mode (SMM) and aligned variable reclaim process (AVRP) to mediate between NVRAM and system memory. This intermediary approach allows fault-tolerant variable reclamation without requiring additional NVRAM capacity, thereby maintaining reliability while improving performance by utilizing existing memory resources more efficiently
Solution Approach 2:
The patent changes the parameter of memory alignment from arbitrary to aligned boundaries, and modifies the reclamation process to use memory-aligned transactions. This parameter change enables efficient use of existing NVRAM without additional capacity requirements, resolving the contradiction between reliability and performance
2Reliability
If additional NVRAM is used for fault-tolerant variable reclamation, then reliability is improved, but cost increases
Solution Approach 1:
The patent uses system management mode as an intermediary layer that enables fault-tolerant reclamation through software-based aligned variable reclaim process, eliminating the need for additional hardware NVRAM capacity and thereby reducing manufacturing costs while maintaining reliability
Solution Approach 2:
The patent implements a copying mechanism where variable data is copied between memory regions during the aligned variable reclaim process. This copying approach enables fault tolerance through data redundancy in system memory without requiring additional NVRAM, thus reducing cost while maintaining reliability
3Device complexity
If conventional variable reclamation is used, then simplicity is maintained, but time consumption increases
Solution Approach 1:
The patent introduces dynamic alignment adjustments in the variable reclaim process, where memory transactions are dynamically aligned based on actual data requirements. This dynamic approach reduces reclamation time by optimizing memory access patterns while maintaining reasonable process complexity through automated alignment handling
Solution Approach 2:
The patent implements continuous aligned variable reclaim operations that proceed without interruption to system operation. The SMM-based AVRP runs continuously in the background, maintaining useful action without stopping system execution, thereby reducing perceived reclamation time while keeping the process integrated and simple
Data Source
AI summary
In one example, reclaiming obsolete regions includes a memory organized in aligned memory blocks and storing valid variables in valid regions and obsolete variables in the obsolete regions. A memory includes a buffer region to cache the memory. A controller can search the buffer region for the obsolete regions and pair with respective valid regions and determine if start addresses of the obsolete regions are memory aligned and if not aligned, to write a small portion content of a first valid region to the start address of the aligned memory block, and to write any remaining respective valid region beginning at the start address of the aligned memory block and in multiples of the aligned memory block. Upon completion of a writing, moved respective valid regions begin at the starting address of the obsolete regions and new obsolete regions begin at end addresses of the moved respective valid regions.