With the release of IBM Db2 Analytics Accelerator (IDAA) Version 8, a powerful new capability has been introduced: the ability to copy data directly between two accelerators. This enhancement significantly simplifies data movement and opens up a range of new use cases for organizations looking to optimize their analytics infrastructure.

Traditionally, moving data between accelerators required a more manual and resource-intensive process, often involving unloading and reloading data. With direct accelerator-to-accelerator data transfer, this process becomes much more efficient, reliable, and scalable.

Let’s explore some of the key use cases enabled by this new feature:

 1. Seamless Migration of Accelerator-Only Tables (AOTs)

One of the most exciting applications of this new feature is the ability to move Accelerator-Only Tables (AOTs) between accelerators. AOTs are a special type of table that exist solely within the accelerator, they are not backed by any data in Db2 for z/OS. These tables are often used to:

• Store data from external sources, such as SMF records or VSAM datasets.
• Serve as in-database transformation targets.
• Hold transient tables that don’t need to reside in the main Db2 for z/OS database.

Because AOT data do not exist in Db2 for z/OS, moving them between accelerators used to be cumbersome, requiring backup and restore operations. Now, with direct copying between accelerators, AOTs can be migrated seamlessly, enabling:

• High availability configurations where AOTs are copied across sites.
• Disaster recovery strategies that include non-Db2 data.
• Simplified data lifecycle management for temporary or external datasets.

This is a special and powerful use case that goes beyond the traditional role of the accelerator as a mirror of Db2 for z/OS data.

2. Cloning Accelerators for Scale-Out and Redundancy

Organizations aiming to scale their analytics capabilities or enhance business continuity can now clone an accelerator more easily, thanks to support for regular shadow tables in the copying process. This enables several strategic deployment scenarios:

• Horizontal scaling: Set up a new accelerator instance using data from an existing one to distribute workloads and expand capacity.
• Disaster recovery readiness: Maintain a synchronized copy of accelerator data in a secondary location to ensure rapid recovery in case of failure.
• Load balancing and failover strategies: Implement redundancy and automatic switchover mechanisms to guarantee uninterrupted analytics services.

This cloning capability is especially valuable in enterprise environments where agility and resilience are essential, whether you're rapidly provisioning test systems or building high availability architectures.

 3. Daisy-Chaining Load Operations to Reduce z/OS Resource Usage

Another innovative use case is the ability to daisy-chain load operations. Instead of repeatedly loading the same data from Db2 for z/OS into multiple accelerators, each time consuming CPU and I/O resources, you can now:

• Load the data once into a primary accelerator.
• Then copy it to additional accelerators as needed.
• This approach significantly reduces the load on the mainframe, freeing up resources for other critical workloads.

This is beneficial in environments with multiple accelerators.

Final Thoughts

The ability to copy data directly between IBM Db2 Analytics Accelerators in Version 8 is a significant advancement that opens up new workload scenarios, such as high availability for Accelerator-Only Tables, faster scale-out deployments, and disaster recovery. At the same time, it reduces operational complexity for maintaining data.

If you're planning to adopt or upgrade to IDAA V8, now is the perfect time to explore how this capability can enhance your data workflow.

Links:
https://www.ibm.com/docs/en/daafz/8.1?topic=tasks-copying-tables-between-accelerators
https://www.ibm.com/docs/en/daafz/8.1?topic=procedures-sysprocaccel-copy-tables