Memory Transfer Database Synchronization Bypassing OS Resources
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing database synchronization methods between active and standby management units in communication networks are resource-intensive and latency-prone, leading to inefficiencies in high-availability operations due to reliance on application-layer processing and network communication.
Innovation Solution
Implementing memory-transfer-based synchronization by tracking updated memory blocks in the active management unit and using direct memory access (DMA) to transfer these blocks to the standby unit, bypassing operating system resources and utilizing a unified memory space for efficient copying.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If application-layer processing and network communication are used for database synchronization, then data consistency is maintained, but resource consumption increases and latency occurs
Solution Approach 1:
The patent segments the database into fixed-size memory blocks and tracks only the updated blocks rather than synchronizing the entire database. This segmentation allows selective transfer of only changed portions, reducing resource consumption while maintaining data consistency for updated elements.
Solution Approach 2:
The patent introduces a synchronization unit as an intermediary layer between the active and standby management units. This intermediary handles the block tracking and transfer coordination, enabling efficient synchronization without requiring full application-layer processing or network communication for each update.
2Reliability
If application-layer processing and network communication are used for database synchronization, then data consistency is maintained, but synchronization latency increases
Solution Approach 1:
The patent performs preliminary actions by pre-allocating memory blocks and establishing block tracking mechanisms before synchronization is needed. The synchronization unit continuously monitors and identifies updated blocks, so when synchronization is required, only the identified blocks need to be transferred, significantly reducing latency.
Solution Approach 2:
The patent extracts the synchronization function from the application layer and implements it at the memory/block level through a dedicated synchronization unit. This extraction eliminates the overhead of application-layer processing and network communication, reducing synchronization latency while maintaining data consistency.
3Use of energy by moving object
If memory blocks are transferred using DMA bypassing operating system resources, then resource utilization decreases, but system complexity increases
Solution Approach 1:
The synchronization unit acts as an intermediary that manages the complexity of DMA operations. It handles block tracking, address mapping, and coordination with the DMA controller, thereby reducing resource utilization while containing system complexity within a dedicated management layer rather than分散 across the entire system.
Data Source
AI summary
A system for memory-transfer-based database synchronization in an appliance is provided. During operation, the system can receive, from a database manager on an active unit, a notification indicating that a first memory block has been updated due to an update to a first entry in a first database of the active unit. The system can identify a second memory block that stores a corresponding second entry in a second database of a standby unit. The first and second memory blocks can be in first and second memory segments, respectively, of respective memory devices of active and standby devices. The first and second memory segments can be outside of the accessible memory range of the respective operating systems. The system can then replace the content of the second memory block with the content of the first memory block using a memory-transfer operation, thereby synchronizing the first and second entries.