Dual-Frequency Data Synchronization for Backup Cost Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data backup and migration techniques in data centers face challenges in ensuring data completeness and response performance while maintaining cost-effectiveness, particularly when multiple backup destinations are used, leading to increased load on the data center and higher costs.
Innovation Solution
The system employs a dual-frequency data synchronization approach, where management data is acquired and updated at a higher frequency, and actual data is synchronized at a lower frequency, allowing for flexible migration and reduced backup costs by utilizing multiple backup destinations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data is backed up to multiple destinations, then data completeness and migration options are improved, but load on data center and costs increase
Solution Approach 1:
The backup system is segmented into two distinct components: management data (metadata, file information, directory structures) and actual data (file contents). This segmentation allows different backup frequencies and storage strategies for each type, reducing overall backup costs while maintaining data completeness.
Solution Approach 2:
The system implements periodic backup actions with different frequencies: management data is backed up at a first frequency (more frequently) while actual data is backed up at a second frequency (less frequently). This periodic action with varying frequencies optimizes the balance between data completeness and backup costs.
2Reliability
If perfect synchronization mirroring is used, then data completeness is ensured, but system response performance decreases
Solution Approach 1:
Instead of continuous real-time synchronization, the system uses periodic backup actions at specified frequencies. Management data is synchronized periodically at a higher frequency, while actual data is synchronized periodically at a lower frequency, reducing the performance overhead while maintaining data completeness.
Solution Approach 2:
The system performs partial synchronization by separating management data and actual data. Only management data is synchronized at high frequency, while actual data synchronization is performed less frequently. This partial action approach ensures data completeness without the full performance cost of complete real-time synchronization.
3Speed
If asynchronous mirroring is used, then system response performance is improved, but data completeness cannot be guaranteed
Solution Approach 1:
The system segments data into management data and actual data, applying different synchronization strategies to each. Management data is synchronized with higher frequency to guarantee completeness, while actual data uses lower frequency synchronization. This segmentation allows the system to achieve both performance and reliability goals.
Solution Approach 2:
The system implements feedback mechanisms through periodic backup operations that check and update the state of both management data and actual data. This feedback ensures that data completeness is maintained through regular synchronization while allowing asynchronous operations between backups to improve response performance.
Data Source
AI summary
A metadata processing unit acquires the first metadata from a first data storing unit at a first frequency and makes a second data storing unit store the first metadata. A file data processing unit specifies, based on the update history of the first metadata, second metadata that is updated on and after the last acquisition date and time of second file data stored in the second data storing unit, acquires the first file data associated with the specified management data at a second frequency that is lower than the first frequency, and makes the second data storing unit store the first file data. A switching unit switches a read/write process that is performed by using the operation data stored in the first data storing unit such that the read/write process is performed by using the second metadata and the second file data in the second data storing unit.