Replication Sub-Clusters for Continuous Data Protection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data protection systems face challenges in maintaining continuous data protection, particularly in keeping pace with high data transaction rates at production sites without causing downtime or data backlog, due to the overhead of journaling processes.
Innovation Solution
The implementation of a method and system that creates sub-clusters at production and replication sites, using replication appliances to split IOs and manage virtual machines, allowing for dynamic adaptation to changing data transaction rates and enabling efficient data replication and recovery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If journaling is used to enable continuous data protection and rollback to any point in time, then data recovery capability is improved, but system overhead increases and may cause production site slowdown
Solution Approach 1:
The system divides the replication appliances into sub-clusters (first sub-cluster at production site, second sub-cluster at replication site) that can independently handle different portions of the data replication workload. This segmentation allows the system to manage journaling overhead more effectively by distributing it across multiple independent units rather than having a single bottlenecked system handle all journaling operations.
Solution Approach 2:
The replication system dynamically adapts to changing data transaction rates by allowing sub-clusters to scale and adjust their capacity. When transaction rates increase, additional replication appliances can be added to the sub-clusters to handle the increased journaling load without forcing the production site to slow down, thus maintaining both reliability and performance.
2Reliability
If backup site keeps pace with high data transaction rates, then data protection effectiveness is improved, but backup site resources and capacity requirements increase
Solution Approach 1:
The backup/replication site is divided into sub-clusters that can be independently scaled. Each sub-cluster handles a portion of the replication workload, allowing the system to incrementally add resources to match growing data protection needs rather than requiring a large monolithic backup infrastructure from the start.
Solution Approach 2:
The replication appliances in the sub-clusters are designed to perform multiple functions: they handle both the journaling operations for continuous data protection and the actual data replication. This multi-functionality reduces the need for separate dedicated resources at the backup site, as the same appliances that replicate data also manage the journaling overhead.
3Stability of the object's composition
If production site is forced to slow down to allow backup completion, then data consistency is improved, but productivity and throughput decrease
Solution Approach 1:
The system performs preliminary journaling actions at the production site without forcing the production workload to slow down. By capturing change data in journals asynchronously and using sub-clusters to process these journals, the system maintains data consistency while allowing the production site to continue operating at full speed. The consistency is achieved through the journal mechanism rather than by throttling production transactions.
Data Source
AI summary
A computer implemented method, system, and computer program product for replicating multiple tenants comprising creating a first sub cluster at a production site; wherein the production site has a cluster of replication appliances and wherein the sub-cluster comprises a subset of the replication appliances in the cluster of replication appliance at the production site, creating a second sub cluster at a replication site comprising of a subset of replication appliances in the cluster of replication appliance at the production cluster, and creating a replication subsystem comprising of the first and second sub-clusters and replicating from the production sub-cluster to the replica sub cluster.