Backup Agent Segregation for Clustered Data Conflict Resolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional backup systems face conflicts and failures when attempting to concurrently back up both clustered and non-clustered data, as they lack the ability to differentiate between data types and manage backup requests effectively, leading to unnecessary duplication and errors.
Innovation Solution
The method involves segregating data into clustered and non-clustered types, allowing concurrent backup of standalone data on individual nodes while queuing backup requests for clustered data until active backup processes are completed, ensuring sequential processing and avoiding conflicts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If concurrent backup processes are attempted for both clustered and non-clustered data, then backup throughput is improved, but backup reliability deteriorates due to conflicts and failures
Solution Approach 1:
The backup system segments backup operations into two distinct types: clustered data backup and non-clustered data backup. This segmentation allows each type to be handled by specialized backup agents that understand the specific requirements and conflict resolution strategies for that data type, thereby maintaining reliability while enabling concurrent operations where safe.
Solution Approach 2:
The backup manager acts as an intermediary that receives backup requests, determines the data type (clustered or non-clustered), and routes requests to appropriate backup agents. This intermediary coordinates between multiple backup agents, manages queuing for clustered data, and enables concurrent execution for non-clustered data, resolving the contradiction between throughput and reliability.
2Reliability
If backup requests are queued sequentially for clustered data, then backup reliability is improved by avoiding conflicts, but backup productivity deteriorates due to waiting time
Solution Approach 1:
The system merges the backup operations of multiple nodes into a single coordinated process for clustered data. Instead of each node independently backing up clustered data (which would cause conflicts), the backup manager consolidates these requests and executes a single backup operation that serves all nodes, maintaining reliability while improving overall productivity by eliminating redundant work.
Solution Approach 2:
The backup manager performs preliminary actions by receiving and analyzing backup requests before execution. It determines data types, identifies clustered versus non-clustered data, and prepares appropriate queuing or concurrent execution strategies in advance, allowing the system to optimize between reliability and productivity based on the specific data characteristics.
3Productivity
If backup requests are processed concurrently for non-clustered data, then backup productivity is improved, but device complexity increases due to coordination requirements
Solution Approach 1:
The system applies local quality by treating different data types differently: non-clustered data on each node is backed up independently without coordination overhead, while clustered data receives coordinated management. This localized approach to backup management reduces overall system complexity by minimizing coordination requirements where they are not needed, while maintaining productivity through concurrent execution of independent backup operations.
Data Source
AI summary
A method for backing up standalone data and clustered data on a node, that includes receiving, by a backup agent, a first backup request, making a first determination that a backup process is not active, and based on the first determination, identifying a first type of data specified in the first backup request, where the first type of data is the standalone data, locating, using a first backup policy, the standalone data in the node, begin generating standalone backup data from the standalone data, receiving, by the backup agent, a second backup request, making a second determination that the backup process is active, and based on the second determination, queuing the second backup request.