ABN AMRO embraces an Azure-first data strategy to drive better business decisions

Big bank meets big data

ABN AMRO needed more from its on-premises servers. Having built a reliable, secure system capable of handling millions of daily transactions, the Dutch bank knew it needed a quicker and more reliable way for its various teams to contribute and access data.

According to Piethein Strengholt, ABN AMRO’s Principal Data Architect, the types of data a bank processes in a typical day are diverse and complex. “We have applications for business process management, HR systems, financial systems, customer systems, marketing systems, and commercial systems. And we record conversations and correspondence with our clients. Those are just some of the types of data we process and collect.”

That’s why the engineering team responsible for managing ABN AMRO’s data turned to Microsoft Azure. Migrating to the cloud meant greater scalability, ease of access, and taking advantage of a suite of Azure services to empower both customers and functional teams alike.

When it came to designing a new architecture, readability of data was of utmost importance to Strengholt. The evolution of analytics has resulted in changes in the read-versus-write ratio. For example, analytical models that are constantly retrained read huge volumes of data. Facilitating this high variety of read patterns while duplicating data and staying in control is a challenge. ABN AMRO found that Azure Data Lake Storage and Azure Cosmos DB allowed it to meet this challenge and scale.

Turning data into value

Accessing data for actionable intelligence is crucial to operational efficiency, but different business domains need to access the same data sets for different purposes. While Azure Data Lake Storage is an optimal solution for storing both structured and unstructured data, a cloud-distributed data mesh offers even more flexibility when managing the needs of  cross-functional teams. ABN AMRO can eliminate silos, while allowing demanding analytical workloads to continuously read data.

The company uses Azure Data Lake Storage and the Parquet file format to make data readily available to a wide range of Azure services, including Azure Databricks and Azure Synapse. This allows Strengholt’s team to access data without having to process it into different formats.

“We embraced a data mesh architecture on Azure, where you see a lot of forward thinking regarding data ownership and security,”

Simply put, a data mesh architecture embraces distributed data using domain-driven design principles. It supports data democratization and “data as a product,” with specific domains managing their own data pipelines. A universal interoperability layer that employs the same syntax and data standards connects these domains and their associated data.

For ABN AMRO, the advantages of using a data mesh architecture with Azure Data Lake Storage were clear from the beginning. When data is coming from multiple sources, in multiple formats, having one logical place to store and distribute that data is the starting point to develop a more responsive, customized solution. A data mesh opens up new possibilities for analysts looking to not just store data, but to use it to drive better business decisions among different teams with different needs.

“We use Azure API Management and Azure Data Lake Storage,” Strengholt says. “We use Azure Databricks for processing data and are looking into Azure Synapse for consuming the data, and we use Azure Data Factory, of course, for orchestrating all the data.”

ABN AMRO’s custom data mesh architecture meant evolving from the bank’s tightly coupled on-premises systems to a cloud-based platform that decoupled compute from storage. This resulted in a vastly improved capacity to scale. Processes that once kept on-premises servers busy for months now take mere days in the cloud.

Better accountability

ABN AMRO used a data mesh architecture on Azure to model and govern data faster and  more accurately. Since this data includes the sensitive financial information of millions of customers, the systems managing it must absolutely meet the most rigorous security standards. But ensuring that an organization meets its regulatory compliance obligations doesn’t mean it has to sacrifice ease of access and speed.

Strengholt elaborates, “We have established a data governance body within our company, and no data is allowed to be distributed or consumed without clear ownership. So, for each data set we onboard on the central platform, we want to ensure data accountability.” 

Azure solutions mean money in the bank

ABN AMRO’s commitment to data security was matched by Azure-native security features, including Azure Active Directory, an enterprise identity service that provides single sign-on and multifactor authentication. This was complemented on the analytics side by Azure Synapse, with a robust set of security features that include dynamic data masking and row-level security. To Strengholt, these features were a game-changer.

Providing ease of access, security, and reliability of data is just the beginning of managing a trusted financial institution like ABN AMRO. Taking banking to the next level means analyzing that data in ways that drive improved insights and smarter business decisions.

Migrating to Azure allowed the bank to capitalize on cloud-native services designed to drive both growth and digital transformation.

Azure API management streamlines work across hybrid and multi-cloud platforms, which is crucial to a bank that still needs to retain some data on-premises and interacts with other cloud platforms.

Ingesting, preparing, and transforming data at scale is what Azure Data Factory was built for. This fully managed, serverless data integration service has automated processes that used to be far more laborious for the legacy systems from which the bank migrated.

With Azure Databricks, ABN AMRO better manages analytics workflows by enabling collaboration, AI insights, and advanced, automated machine learning capabilities. This means that not only are functional teams able to process data more efficiently, they’re also empowered to use that data in ways that simply weren’t possible before. Shared workspaces and interoperation with Azure Synapse Analytics for high-performance data warehousing make for unparalleled levels of performance and scalability. 

To analytics and beyond

Ingesting and managing data at scale was just the start of ABN AMRO’s digital transformation strategy. Putting that data to work means taking advantage of the kinds of insights available through Azure Synapse Analytics.

Synapse brings together data integration, enterprise data warehousing, and big data analytics to give users more tools to see the forest through the trees. Analysts at ABN AMRO can now query data on their own terms, using either serverless or dedicated resources — at scale.

All these Azure products are designed to seamlessly work together in precisely the way ABN AMRO needs them to. The company’s engineers and diverse user groups have been duly impressed by the integrated power and future-proofed innovations of Azure offerings.

As for its own growth plans, ABN AMRO anticipates getting even more from Azure. According to Strengholt, “I foresee a future where we have a much smaller on-premises footprint. There’s so much data processing and analytics we can do, making predictions, doing all kinds of complex calculations, and developing entirely new, cutting-edge use cases, for example with Azure Machine Learning Services. All of it can be done on Azure.”