Credit bureau · data replication

Replicating legacy DB2 data into a big-data platform in real time

Implemented production replication and streaming from an AIX DB2 estate into modern DB2, Cloudera Hadoop, and AWS targets.

Delivered engagement
AIX → Red HatDB2 migration
2production environments
DB2to Hadoop and AWS

The challenge

A credit bureau needed to move an AIX-based DB2 production estate onto a Red Hat DB2 environment while also streaming operational changes into a big-data platform and AWS.

The solution required production installation, source-to-target reconciliation, performance coordination, and recoverable operating procedures across legacy and modern platforms.

What Monic delivered

Monic migrated DB2 11.1 workloads from AIX to Red Hat, installed and validated two production Attunity Replicate 6.5 environments, and configured data streaming into Cloudera Distribution Hadoop 6.0.1 and AWS targets.

The team defined source-table requirements, designed reconciliation methods, worked with database administrators on replication performance, troubleshot migration and streaming issues, and documented installation and disaster-recovery procedures.

The result

The three-month implementation established production replication paths from the legacy DB2 estate into the target database and big-data environments, with validation, reconciliation, and recovery practices documented for ongoing operation.

What this case demonstrates

The engineering lessons behind the result.

These are the delivery patterns a team facing a similar operating problem should plan for.

One source can serve several targets

A replication design can coordinate a database migration, big-data ingestion, and cloud delivery when each target has explicit requirements and validation.

Reconciliation is cross-team work

Source definitions, target checks, and performance tuning require coordination between replication engineers, database administrators, and downstream platform owners.

Recovery documentation is a deliverable

Installation records, operating procedures, and disaster-recovery steps reduce dependence on implementation memory after production handover.