Splitter-Based Data Replication for High-Transaction Backups
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data protection systems face challenges in minimizing downtime during recovery and ensuring continuous data protection, as they struggle to keep pace with high data transaction rates at production sites, leading to potential system shutdowns due to backlog and inefficiencies in journaling processes.
Innovation Solution
A computer-implemented method involving a splitter that intercepts IO data, sends it to a replication protection appliance (RPA), and then to a remote site, with metadata and timestamps managed to ensure efficient replication and synchronization, allowing for dynamic adaptation to changing transaction rates and minimizing downtime.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If continuous data protection is implemented using journaling to enable rollback to any specified point in time, then recovery capability is improved, but the backup site cannot keep pace with high data transaction rates at the production site, causing backlog and potential system shutdowns
Solution Approach 1:
The patent introduces a splitter as an intermediary component between the production site and backup site. The splitter intercepts I/O requests, sends metadata to the production RPA for processing, and only transfers data to the backup site after receiving acknowledgment. This mediator architecture allows the backup site to keep pace with production transactions without forcing production to slow down, resolving the contradiction between recovery capability and transaction processing rate.
Solution Approach 2:
The patent implements preliminary action by sending metadata to the production RPA before transferring actual data to the backup site. The production RPA processes and acknowledges the metadata first, ensuring that the backup site is ready to receive data. This preliminary processing of transaction information allows the system to maintain synchronization without backlog, enabling continuous protection at high transaction rates.
2Reliability
If synchronous replication is used to ensure data consistency at the backup site, then data protection reliability is improved, but downtime during recovery increases and productivity decreases
Solution Approach 1:
The patent implements dynamic replication by allowing the system to adapt its behavior based on real-time conditions. The splitter dynamically sends metadata to the production RPA and waits for acknowledgment before transferring data. This dynamic approach ensures data consistency is maintained only when necessary, allowing faster recovery by not waiting for synchronous completion of all backup operations, thus reducing recovery downtime while maintaining reliability.
3Reliability
If the backup site processes every data transaction to maintain continuous protection, then data protection is improved, but the overhead of journaling requires several data transactions at the backup site for each production site transaction, slowing down the production site
Solution Approach 1:
The patent extracts only the essential information (metadata) from production transactions and sends it to the backup site through the splitter. Instead of transferring complete data transactions, the system extracts and transmits only the necessary transaction metadata, reducing the overhead at the backup site. This extraction approach maintains continuous protection capability while significantly reducing the number of transactions the backup site must process, thereby preventing production site slowdown.
Solution Approach 2:
The patent applies partial action by processing only the metadata portion of transactions at the backup site rather than complete data transactions. The backup site processes metadata to maintain protection records but does not need to fully process every data transaction, reducing the workload multiplier effect. This partial processing approach maintains sufficient protection while reducing the overhead that would otherwise slow down production transactions.
Data Source
AI summary
A computer implemented method, a computer program product, and a system for replicating data comprising: intercepting the IO data to a volume at the splitter; sending the IO data intercepted at the splitter to the volume; sending metadata of the IO data to a production RPA from the splitter, wherein an acknowledgement is sent to the splitter indicating that the metadata was received by the production RPA; and sending IO data, metadata, and timestamp of IO data to a remote site from the splitter, wherein IO data is sent to the remote site after splitter receives acknowledge of production RPA receiving the IO data.