Parallel Partitioned Data Container Backup via Snapshot Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current backup processes for large data containers, such as virtual machines, are inefficient due to serial data reading and high resource wastage, leading to long backup windows and network latency issues.
Innovation Solution
A system and method that generates a snapshot image of a data container, allowing simultaneous data reading from multiple partitions using multiple threads over transport paths, ensuring the data retains its physical block sequence and is backed up efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If serial data reading is used to backup data containers, then resource usage is reduced, but backup throughput decreases and backup time increases
Solution Approach 1:
The patent divides the data container into multiple partitions and assigns each partition to a separate thread for parallel reading. This segmentation allows simultaneous access to different portions of the data container, thereby increasing backup throughput and reducing overall backup time without overwhelming system resources.
Solution Approach 2:
The patent transitions from single-threaded serial reading to multi-threaded parallel reading by adding the dimension of concurrency. Multiple threads operate simultaneously on different partitions, effectively utilizing available I/O bandwidth and processor capacity to improve backup performance.
2Productivity
If multiple threads are used for simultaneous data reading, then backup throughput increases, but system resource consumption increases
Solution Approach 1:
The patent uses multiple threads to read from different partitions simultaneously, applying partial parallelism rather than full system-wide parallelism. This approach increases throughput by utilizing available I/O capacity without requiring proportional increases in overall system resources, as each thread operates on a specific partition independently.
3Speed
If data is read simultaneously from multiple partitions, then backup speed increases, but data sequence integrity may be compromised
Solution Approach 1:
The patent implements a coordination mechanism where multiple reading threads report their progress and the system tracks the physical block sequences being read. This feedback allows the system to maintain proper sequencing of data blocks during parallel reading operations, ensuring integrity while achieving speed improvements through concurrency.
Data Source
AI summary
A method for improving performance of data container backups comprises identifying a data container on a source computer to be backed up. A snapshot image of the data container is generated, wherein the snapshot image comprises one or more partitions. Data is read simultaneously from each of the one or more partitions to a destination computer over one or more transport paths. Each of the one or more transport paths comprises one or more threads. The data read to the destination computer is backed up such that the data retains a physical block sequence of the data container.