In the digital banking era, data is the cornerstone of innovation, enabling personalized services, regulatory compliance, and operational efficiency. Banks must manage vast datasets, share them seamlessly with internal teams and external partners, and maintain robust security. Three global banks—ABN AMRO, ING, and HSBC—have addressed these challenges with innovative data integration architectures: ABN AMRO’s Digital Integration and Access Layer (DIAL), ING’s cloud-based data lake, and HSBC’s API-led data mesh. Despite their differences, these approaches share a focus on flexibility, real-time connectivity, and ecosystem collaboration.
This article explores each bank’s strategy, key features, outcomes, and lessons for the banking sector, supported by a comparative table. The approaches of the three Banks outline multiple pathways for streamlining data architecture.
Case Studies
ABN AMRO: Digital Integration and Access Layer (DIAL)
Why Change Was Needed
ABN AMRO’s legacy data system, a centralized hub, struggled to meet modern banking demands. It was slow to process real-time updates, integrate external data sources, and comply with evolving regulations like GDPR. The bank needed a scalable, flexible architecture to support digital services and partnerships with fintechs, prompting a complete overhaul of its data management approach.
The Solution: DIAL
ABN AMRO developed the Digital Integration and Access Layer (DIAL), a centralized middleware that orchestrates data flows across systems. DIAL ensures data is consistent, secure, and accessible while minimizing redundant transformations. It leverages metadata—labels detailing data’s meaning, origin, ownership, and quality—to build trust without enforcing a rigid, universal data model. This metadata-driven approach simplifies data discovery and ensures compliance with regulatory requirements.
How It Works
DIAL supports multiple data movement methods tailored to specific needs. Read Data Stores (RDS) provide read-only data copies, protecting source systems from overload. APIs enable real-time communication between applications, while event-driven architecture delivers instant updates via event notifications, ensuring systems stay synchronized. Data reaches Domain Data Stores (DDS), customized for specific business units like lending or customer analytics. Valuable new data is registered in the List of Golden Sources (LoGS), a catalog of trusted sources for enterprise-wide reuse. DIAL fosters an ecosystem by connecting internal systems with external partners, defining clear roles: data providers supply and manage data, while data consumers use it responsibly. This ecosystem approach supports integrations with fintechs for services like instant payments or open banking APIs mandated by PSD2.
Outcomes
DIAL has accelerated data access and processing, enabling faster development of customer-facing services like mobile banking apps. It has strengthened partnerships with fintechs through seamless data sharing, enhancing offerings like real-time financial insights. Metadata and lineage tracking have improved governance, ensuring compliance with regulations and reducing audit times. While still evolving (based on 2018 insights), DIAL positions ABN AMRO as a leader in data-driven banking.
ING: Data Lake and API-Driven Architecture
Why Change Was Needed
ING’s siloed legacy systems created bottlenecks, delaying innovation and hampering customer service. The bank needed a unified, scalable platform to manage diverse data types—transactional, customer, and market data—while supporting digital transformation and partnerships with fintechs and cloud providers. A modern architecture was essential to stay competitive.
The Solution: Data Lake and APIs
ING implemented a cloud-based data lake on Microsoft Azure to centralize structured and unstructured data from legacy systems. Complemented by an API-driven platform, this architecture enables real-time data access and integration with external partners, mirroring DIAL’s ecosystem focus. The data lake serves as a single source of truth, supporting analytics, machine learning, and customer applications.
How It Works
The data lake consolidates data from across ING’s operations, making it accessible for business intelligence and personalized services. APIs connect internal systems and external partners, powering services like instant payments, account aggregation, and personalized loan offers. Event streaming ensures real-time updates, keeping systems aligned with customer interactions. Metadata management—similar to DIAL’s approach—tracks data quality, lineage, and ownership, ensuring governance and compliance with regulations like PSD2. ING’s architecture supports advanced analytics, such as predictive models for fraud detection, and integrates with cloud services for scalability. The bank’s DevOps culture accelerates deployment, with automated pipelines for data ingestion and API development.
Outcomes
ING reduced time-to-market for new digital services , launching features like mobile payment solutions faster. Customer engagement rose driven by personalized experiences, such as tailored savings recommendations. Centralized governance and metadata improved regulatory compliance, streamlining audits. The data lake’s scalability supports ING’s global operations, making it a robust counterpart to ABN AMRO’s DIAL.
HSBC: API-Led Integration and Data Mesh
Why Change Was Needed
HSBC’s global operations relied on legacy mainframes that couldn’t support real-time, cross-border data sharing. The bank needed a scalable, decentralized architecture to connect its diverse systems, enable digital services like instant loan approvals, and comply with global regulations such as GDPR and PSD2. A modern integration strategy was critical.
The Solution: API-Led Data Mesh
HSBC adopted MuleSoft’s API-led integration and a data mesh architecture to decentralize data ownership while ensuring interoperability. Like DIAL’s provider-consumer model, this approach assigns clear responsibilities to business domains. The data mesh empowers domains to manage their data, sharing it via APIs, while API-led integration connects legacy and cloud systems.
How It Works
HSBC’s API platform uses three layers: System APIs access legacy data, Process APIs orchestrate business logic, and Experience APIs deliver tailored user interfaces. This layered approach, similar to DIAL’s middleware role, bridges old and new systems. Kafka-based event streaming supports real-time updates, aligning with DIAL’s event-driven flows. The data mesh decentralizes data management, allowing domains like retail banking or wealth management to own their data while sharing it via standardized APIs. Metadata ensures security, compliance, and lineage, enabling traceability across HSBC’s global network. The architecture supports open banking initiatives, integrating with fintechs for services like payment gateways.
Outcomes
HSBC cut integration time for new services , accelerating rollouts of features like real-time account management. Customer satisfaction increased through personalized, real-time services, such as instant credit approvals. Metadata-driven governance ensured compliance with global regulations, reducing regulatory risks. The decentralized data mesh suits HSBC’s global scale, distinguishing it from DIAL’s centralized design.
Comparison of Architectures
Lessons for the Banking Sector
These case studies reveal shared principles and unique insights for modern banking.
First, embracing ecosystems is non-negotiable. Architectures like DIAL, data lakes, and data meshes enable seamless collaboration with fintechs, cloud providers, and regulators, supporting open banking and customer-centric services.
Second, metadata is a cornerstone, providing trust, compliance, and usability without rigid, monolithic data models.
Third,agility drives success—real-time APIs and event-driven systems accelerate innovation, from instant payments to personalized offers.
Finally, extent/level ofcentralization must align with scale. ABN AMRO’s centralized DIAL suits its regional focus, ING’s semi-centralized data lake balances global and local needs, and HSBC’s decentralized data mesh supports its vast, global operations. Banks must choose a model that fits their scope while prioritizing flexibility and governance.
The differences in these architectures highlight strategic trade-offs. ABN AMRO’s custom-built DIAL offers tight control but requires ongoing investment to scale. ING’s Azure-based data lake leverages cloud scalability but demands robust governance to prevent data sprawl. HSBC’s data mesh excels in decentralized environments but introduces complexity in coordinating domain-level data. Despite these variations, all three banks demonstrate that modern data integration is about reimagining data flows to create value, ensure compliance, and delight customers in a digital-first world.
( Based on publicly available information. It is difficult to always access upto date information, so it is possible the case studies at times do not reflect current level of data architecture)
Glossary
DIAL (Digital Integration and Access Layer): ABN AMRO’s central system for managing data sharing, access, and transformation.
Data Lake: ING’s centralized repository for structured and unstructured data, hosted on Azure.
Data Mesh: HSBC’s decentralized approach where business domains manage and share data via APIs.
Data Provider: Systems that produce or own data (used in all three cases).
Data Consumer: Systems or applications that use data (used in all three cases).
Read Data Store (RDS): ABN AMRO’s read-only data copies to prevent overloading source systems.
Domain Data Store (DDS): ABN AMRO’s tailored data storage for specific business needs.
API (Application Programming Interface): Used by all three banks for real-time system communication.
Event-Driven Architecture: Real-time updates via events, used across all three architectures.
Metadata: Data labels describing meaning, origin, and quality, critical for governance in all cases.
Lineage: Tracking data’s origin and transformations, used for compliance.
LoGS (List of Golden Sources): ABN AMRO’s list of trusted data sources.
Integration Architecture: The overall setup for data sharing and management, customized by each bank.
Leave a Reply