Data Replication Block Ordering via Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data replication methods are inefficient due to their inability to adapt to changing bandwidth and typically utilize only a single replication link, leading to prolonged replication times.
Innovation Solution
A system that divides data into blocks and determines the optimal order for transmission using a machine learning method combining Bagging with Random Forest and ontology evaluation algorithms, leveraging multiple available replication links to minimize replication time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional data replication methods are used, then the replication process is simple to implement, but the replication time is prolonged and the system cannot adapt to changing bandwidth conditions
Solution Approach 1:
The patent implements dynamic adaptation by continuously monitoring bandwidth conditions across multiple replication links and adjusting the replication strategy in real-time. The system transitions from static, single-link replication to dynamic, multi-link replication with adaptive block ordering, allowing the replication process to respond to changing network conditions and minimize replication time
Solution Approach 2:
The patent divides the data into multiple blocks and uses a machine learning model to determine the optimal transmission order of these blocks across multiple replication links. This segmentation allows parallel transmission of different blocks over different links, improving overall replication speed and adaptability to varying bandwidth conditions
2Productivity
If a single replication link is used, then the system complexity is low, but the replication efficiency is reduced
Solution Approach 1:
The patent combines multiple replication links into a unified replication system, allowing simultaneous use of multiple network paths. By merging the capabilities of multiple links and coordinating them through a block ordering model, the system achieves higher replication efficiency while managing complexity through structured coordination
Solution Approach 2:
The patent introduces a machine learning-based block ordering model as an intermediary between the data source and replication targets. This intermediary optimizes the transmission sequence of data blocks across multiple links, enabling efficient use of multiple replication paths without requiring complex coordination logic at the network level
Data Source
AI summary
A processor determines the data to be replicated in the production site. The processor splits the data into a plurality of blocks. The processor determines one or more replication links. The processor determines an order of each one of the plurality of blocks to be sent over the one or more replication links, where the order of each one of the plurality of blocks to be sent over the one or more replication links is determined by a machine learning model.