Replicated Data Consistency Verification via Chunk Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data protection systems lack effective methods for continuous and thorough consistency verification of replicated data, relying on inefficient methods like the 'boot test' that do not guarantee data identity at a specific point in time and require generating a working machine, which is invasive and not continuous.
Innovation Solution
A computer-implemented method and system that uses a target disk and journal to verify data consistency by dividing the protected storage system into chunks, generating a replica of selected chunks at a specified time point, and comparing the copied data with the replica, allowing for continuous online verification without the need to generate a working machine.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the boot test method is used to verify consistency of replicated data, then verification can be performed, but it requires generating a working machine which is invasive and does not enable continuous online verification
Solution Approach 1:
The patent creates a copy of the protected system at the recovery site and verifies consistency by comparing data between the protected system and the copied system without needing to boot the copied system as a working machine. This allows verification without system intrusion while maintaining reliability.
Solution Approach 2:
The patent replaces the mechanical boot process with a data comparison mechanism. Instead of booting the recovered system and checking if it runs, the system directly compares data blocks between the protected and recovered systems, substituting a mechanical operation with a data-level verification process.
2Reliability
If the boot test method is used, then consistency verification is possible, but it is inefficient and cannot be performed continuously online
Solution Approach 1:
The patent enables continuous verification by performing data comparisons online without interrupting system operations. The verification process runs continuously in the background, comparing data blocks as they are replicated, rather than requiring periodic boot tests that interrupt service.
Solution Approach 2:
The patent performs verification actions preliminarily by checking data consistency immediately after replication occurs, rather than waiting for a boot test. This preliminary verification ensures data integrity is maintained throughout the replication process, improving overall efficiency.
3Reliability
If the boot test method is used, then verification can be performed, but it only verifies boot process completion and not data identity at specific time points
Solution Approach 1:
The patent divides the verification process into segments by comparing individual data blocks or chunks between the protected and recovered systems at specific time points. This segmentation allows precise verification of data identity at specific moments rather than only verifying overall system boot success.
Solution Approach 2:
The patent replaces the mechanical boot process verification with direct data comparison at the block level. This substitution enables precise measurement of data consistency at specific time points by comparing actual data content rather than merely checking if the system boots.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system and method for verifying consistency of data in a recovery system, the method including: protecting data in the recovery system, the recovery system includes a target disk configured to store a replica of a protected storage system from before a time window and a journal including data and metadata of write operations to the protected storage system that occurred during the time window; verifying consistency of the replicated data, concurrently with the protecting by: copying data of a selected chunk of the protected storage system at a specified time point, while monitoring write operations to the protected system; generating a replica of the selected chunk in the recovery system from the target disk and the journal, from the specified time point; comparing the copied data of the selected chunk with the replica of the selected chunk and repeating the copying, generating and comparing for a plurality of chunks.