Legal & General increases safety, security, and business agility with Azure SQL DB

Legal and General is based in London, United Kingdon with 1.196 trillion GBP(2022) assets under management and total assets of 513.3 billion GBP(2022). They provide solutions in investments, insurance, workplace pensions and retirement income. They are a leading financial services group and major global investor, safeguarding people’s financial futures around the world. Legal & General’s commitment to prioritize their digital transformation meant that they needed an agile infrastructure that exceeds current industry standards. 

L&G’s existing IT infrastructure needs to evolve with the new business needs and models. So, modernizing their existing data center was crucial to maintaining their operational excellence. Azure’s cloud native managed services helped them handle the increase in data volume and enable them to focus on their business priorities. 

Innovation at enterprise scale

Asset & Liability Payment Service (ALPS) is the core application for handling their retirement policies and retail pension businesses. It is a highly transactional system with 1.5 million lines of code in Java and servicing 800,000 customers. This application had to be modernized from a monolithic architecture to a distributed microservices based architecture. The transactional nature of the workload meant that the system was highly active during certain hours of the day or during certain days of the week. Therefore, the on-premises Db2 database had to be provisioned for peak performance even when such performance was not needed. Alongside this, as new customers get added, the database needs to be scaled, aligning to the increase in workload. "Safeguarding customer systems with a firm and modern technical foundation, addressing current software limitations that risk destabilizing or increasing risk was our primary objective," says Legal & General Senior Programme Manager, Alex McArthur. Modernizing to Azure SQL and Azure Data services solves the problem of scalability, reduces complexity, and fits effectively into the microservices architecture. L&G’s core goals with this modernization were safety, security, resilience and recovery.

ALPS was running on legacy platforms which was difficult to maintain from both technical skills and technology perspective. This modernization enabled L&G to foster agility in introducing new functionality and maintaining existing application. Though cost optimization was not a primary objective for this modernization, L&G noticed that their TCO was reduced by 50% on Azure. 

Modernizing the data landscape to SQL

After a comprehensive assessment of the existing Db2 database, L&G identified critical requirements for any target database including scalability, high availability, disaster recovery within 3 hours, manageability, maintenance flexibility and migration efficiency. Considering all these factors, Azure SQL Database was selected because of its salient capabilities like flexible maintenance windows, geo-replication, optimized locking, database monitoring, automatic tuning, and versionless SQL.

Additionally, Azure SQL provides built-in scalability, high availability, accelerated database recovery within 3 hours, and advanced security features like transparent data encryption using customer managed key. It scales with data and comes with built in intelligence to monitor the cost and on the telemetry data collected. 

A comprehensive migration toolkit 

SQL Server Migration Assistant (SSMA) for Db2 was the tool that was used to assess the conversion gap and then to migrate the schemas and objects from Db2 to Azure SQL Database. The SSMA reports highlighted the conversion compatibility between Db2 and SQL Database while also pointing out areas where automatic conversion was not recommended. The tool generated the report for 3500 objects in an hour and provided 99.9% automatic conversion to SQL.

SSMA for Db2 handled the conversion of tables, indexes, triggers, functions, stored procedures, aliases, synonyms, sequences, and views. The flexibility to modify the data types mapping based on the system requirements enabled L&G to make sure that they got the best possible schema for ALPS. The SSMA tool was used to migrate around 350 million records in under 3 hours from Db2 to SQL Database. 

Data Access Migration Toolkit (DAMT) is a Visual Studio extension that helps to find the SQL Server commands and queries embedded in the application code. The ALPS codebase is Java based with Db2 as backend. The DAMT tool helped to find the embedded SQL queries in the ALPS Java application.

Database Compare Utility is a tool that was used to compare the Db2 and SQL data for post migration validation. The tool verifies the data is identical between source and target. If there is any mismatch it reports in a separate output file for further analysis. For ALPS, this tool compared ~350 million records in 2 hours and validated the data that was migrated to SQL. The tool uses an MD5 hash to compare the hash values between the source and target records. 

Azure Kubernetes Service (AKS) migrates the on-premises application to containers and runs them as fully managed Kubernetes service with AKS. It deploys and manages microservices architecture for their modernized ALPS system. AKS is implemented with Azure DevOps for continuous integration and deployment (CI/CD). 

Azure DevOps provides automated deployment and schedule pipelines at specific times. As the applications are modernized to microservices and containerized, it paves way for flexible deployment strategies. CI/CD approach helps to modernize and expand portfolios continuously by focusing more on business requirements.

An agile data transformation to Azure

By decoupling from their on-prem infrastructure, L&G has been able to build out their Azure deployment in preparation for a well-defined microservices based architecture on AKS and Azure SQL. This will enable the L&G team to increase their focus on core business and build new functionalities to increase the ROI of the business. 

The L&G data modernization was started by migrating data to Azure SQL. In the future, it could be expanded to Microsoft Dynamics, Power Apps, and Azure Open AI.

Next steps

Incrementally transform the ALPS monolith into micro-services by replacing/moving encapsulated functionalities to new services as per Strangler Fig pattern