Synchronous Data Replication via Partial Volume Acknowledgment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional distributed storage systems face challenges in maintaining data consistency and performance during continuous replication, particularly when dealing with damaged volumes, as they often require acknowledgments from all replica volumes before acknowledging write transactions, leading to increased recovery time objectives and reduced system performance.
Innovation Solution
The system intercepts write transactions from a host and sends them to both the production and copy volumes, acknowledging the transaction once acknowledgments are received from a predetermined number (N) of copy volumes, allowing for data consistency even if up to N volumes are damaged, thereby reducing the need to wait for all volumes to acknowledge, thus improving performance and robustness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If acknowledgments are required from all replica volumes before acknowledging write transactions, then data consistency is maintained, but system performance deteriorates and recovery time increases
Solution Approach 1:
The patent applies partial action by requiring acknowledgments from only N out of M replica volumes (where N < M) before acknowledging write transactions to the host. This partial acknowledgment approach maintains data consistency while improving system performance by reducing the number of acknowledgments needed, thus resolving the contradiction between reliability and productivity.
2Reliability
If acknowledgments are required from all replica volumes, then complete data redundancy is ensured, but recovery time objective increases
Solution Approach 1:
The patent reduces recovery time by requiring acknowledgments from only N replica volumes instead of all M volumes. This partial acknowledgment strategy ensures sufficient data redundancy for recovery while minimizing the time loss associated with waiting for all replicas to confirm, thus resolving the contradiction between data redundancy and recovery time objective.
3Reliability
If write transactions wait for all volume acknowledgments, then data integrity is maintained, but system robustness deteriorates
Solution Approach 1:
The patent enhances system robustness by allowing write transactions to proceed after receiving acknowledgments from N volumes rather than waiting for all M volumes. This partial acknowledgment approach maintains data integrity while improving system robustness against volume failures and network issues, resolving the contradiction between data integrity and system adaptability.
Data Source
AI summary
Described embodiments provide systems and methods for operating a storage system. One or more production volumes of the storage system are selected for continuous replication. A number, N, is selected that is associated with a number of damaged volumes the storage system can sustain and maintain data consistency. Write transactions from a host to an associated one of the selected one or more production volumes are intercepted. The intercepted write transactions are sent to the associated production volume and to a plurality of copy volumes. When acknowledgments of the write transaction have been received from N copy volumes, the write transaction is acknowledged to the host.