Decoupling the Behemoth: How Kafka Powers Bradesco's Real-Time Mainframe DB2 Offload to Cloud (Architecting a Resilient CDC Pipeline with Topics, Schema Registry, and Connectors at One of the World's Largest Banks)

The Core Banking Challenge
Bradesco serves over 70 million customers and processes billions of transactions annually on a Mainframe DB2 core that has been the backbone of Brazilian banking for decades. In today's fast-paced financial market, accelerating the delivery of practical, everyday solutions to customers is not just a differentiator — it is a survival imperative. The structural challenge is profound: how do you democratize access to mission-critical transactional data locked across more than 30,000 tables inside a Mainframe DB2, without compromising the non-negotiable robustness, security, and regulatory compliance of one of Latin America's largest banks?

The Modernization Infrastructure
This session reveals the application modernization blueprint driving the next generation of our financial services platform. We will dive deep into our Change Data Capture (CDC) strategy, demonstrating how we leverage IBM InfoSphere Data Replication to capture and stream DB2 changes in real-time — with strict SLAs on latency and near-zero performance impact on the legacy core. We will share the hard-won lessons on tuning CDC subscriptions at scale, managing schema evolution on a platform where "ALTER TABLE" carries existential risk, and navigating the operational realities of bridging a massive z/OS environment to the cloud.

Kafka as the Strategic Decoupler & Resiliency Buffer
The heart of our architecture is the Kafka ecosystem. Far beyond a simple data pipeline, Kafka acts as the strategic decoupling layer between the Mainframe and Azure Cloud. You will learn how:

• Topics and partitioning strategies serve as a critical resiliency buffer — absorbing traffic spikes during peak banking events (payroll days, tax deadlines, Pix instant payment surges) and guaranteeing zero data loss with exactly-once semantics.
• Schema Registry enforces strict governance and data contracts in motion, enabling controlled schema evolution across thousands of complex entities without breaking downstream consumers.
• Kafka Connect with purpose-built sink connectors seamlessly routes governed data into cloud-native destinations — Azure SQL Server for relational workloads and MongoDB for flexible document-oriented consumption patterns.
• Observability and operational resilience — how we monitor lag, detect anomalies, and handle failover scenarios across a hybrid on-premises/cloud Kafka deployment serving a Tier-1 financial institution.
  
  

The Impact: From Bottleneck to Data as a Product
We will share concrete, measurable outcomes: replication latency targets, the growing percentage of our 30,000+ DB2 tables onboarded, consumer squad adoption rates, and the reduction in time-to-market for new digital products. By transforming a rigid legacy infrastructure bottleneck into an agile, event-driven data mesh enabler, we empower autonomous development squads to consume governed, real-time data as a product — bringing new financial solutions to market with unprecedented speed while meeting the stringent security and regulatory demands of the Brazilian Central Bank (BACEN).

Who Should Attend
Architects, engineers, and technical leaders facing mainframe modernization challenges — especially those evaluating CDC-to-Kafka patterns for regulated industries where failure is not an option.