Use cases for Db2 Warehouse replication
The replication feature in Db2® Warehouse on Cloud Pak for Data, which is based on log capture/transaction replay technology, lets you selectively replicate tables between two or more Db2 Warehouse instances on Cloud Pak for Data asynchronously and regardless of location, with low latency for very high-volume workloads.
Some of the key use cases for this feature include:
- Continuous availability
- Continuous availability is more than high availability or disaster recovery. You can achieve high availability with redundant copies of your databases and more reliable components. For disaster recovery, you can use geographically distant copies, backups, and mirrored disks. Continuous availability, on the other hand, requires applications to be always on, including during planned maintenance, upgrades, migrations, and configuration changes. The replication feature in Db2 Warehouse provides the ability to failover and fallback during planned and unplanned outages. You can protect applications for continuous availability by replicating transactions on the tables that are used by these applications.
- Keeping active copies of data in proximity to users
- Replication enables data locality, for example allowing east coast clients to access data from an east coast data center, and west coast clients from a west coast data center. A configuration of this type must ensure that the same piece of data cannot be updated from either data center (row-level) until it is verified that replication has caught up. That is, the client is responsible for rerouting workloads appropriately for avoiding replication conflicts. To ensure that conflicts can be detected, you must have a unique constraint on each table to uniquely identify each row across all databases where the row might be replicated.
- Compliance/regulatory
- Replication for continuous availability enables your organization to meet regulatory requirements to maintain data copies in sync and conduct failover and fallback on a regular schedule for conformance, without the need for an application outage.
- Development and testing
- You can conduct development or unit testing on replicated copies of your databases, or create test systems for data science use-cases with selective data requirements.