Record Level Multiplexing Backup Data Storage Media
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data backup methods, specifically chunk level multiplexing, lead to network congestion and complicate the recovery process by separating data chunks from different save sets on storage media with chunk headers, making recovery inefficient and resource-intensive.
Innovation Solution
Implementing record level multiplexing, where each save set's data chunks are written to storage media without separation by chunk headers, using multiple circular buffers to ensure chunks from one save set are grouped together, with a single record header identifying the save set, thereby improving both backup and recovery efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If chunk level multiplexing is used to transmit backup data, then network congestion is reduced by packaging save sets into manageable chunks, but the recovery process becomes complicated and inefficient due to separation of data chunks by chunk headers
Solution Approach 1:
The patent segments backup data into chunks for efficient network transmission, but then reassembles them at the media record level rather than maintaining chunk-level separation. This allows efficient backup transmission while simplifying recovery by grouping related chunks together in a single media record with a single record header.
Solution Approach 2:
The patent merges multiple chunks from the same save set into a single media record, combining them under one record header rather than keeping them separate with individual chunk headers. This merging eliminates the need for extensive searching during recovery and reduces processing complexity.
2Measurement precision
If multiple chunk headers are written for each chunk to identify originating save set, then chunk identification is accurate, but storage space is wasted and recovery is slower due to extensive searching required
Solution Approach 1:
The patent creates a single record header that serves multiple functions: it identifies the save set for all chunks in the media record, eliminates the need for individual chunk headers, and provides efficient access during recovery. This universal header replaces multiple redundant chunk headers.
Solution Approach 2:
The patent extracts the save set identification function from individual chunk headers and consolidates it into a single record header at the media record level. This extraction eliminates redundant identification data while maintaining accurate save set identification.
3Productivity
If chunks from different save sets are mixed in the same buffer, then buffer utilization is maximized, but recovery efficiency decreases because chunks must be searched for across multiple media records
Solution Approach 1:
The patent segments the buffer into separate circular buffers for each save set, ensuring that chunks from the same save set are grouped together. This segmentation maintains buffer utilization efficiency while enabling fast recovery by locating all chunks of a save set in a single media record.
Solution Approach 2:
The patent applies different buffering strategies to different save sets, with each save set having its own dedicated circular buffer. This local quality approach ensures that each save set's chunks are kept together locally, facilitating efficient recovery without compromising overall buffer utilization.
Data Source
AI summary
A system and method for efficiently creating a backup volume that is also efficient to recover is described. The backup volume may be created with one save set per media record by utilizing record level multiplexing rather than chunk level multiplexing. Through the use of multiple circular buffers, with each buffer associated with a respective save set, save sets written to storage media are contiguous and individual save set chunks are not separated by chunk headers within the respective media record. Recovery is therefore more efficient since the step of reading chunk headers and locating all the chunks for a save set is no longer required.