SaaS Core Banking Transformation
A core banking suite moved to hybrid multi-cloud, bank by bank.
25+banking clients on the platform
$4T+assets under management of the banks moved
Weeks → hoursto provision a bank’s environment
The problem
The platform ran entirely on a private cloud: 1,700 databases, 30,000 virtual servers and 5,000 physical ones, stitched together with many third parties. At its heart sat a large Oracle-based suite with its business logic tied to the data, and earlier attempts to move it to cloud databases had failed.
The answer was to split it. Keep the core data where regulation and performance need it, and move everything around it to each bank’s cloud of choice.
The idea
1Keep the coreThe data layer on Oracle’s dedicated cloud region, inside the client’s walls.
2Move the edgeServices that face users, in containers, to Azure, AWS, Google Cloud or Oracle.
3One modelA unified data model that hides how each cloud names and shapes things.
4AutomateEvery environment built from blueprints, in hours instead of weeks.
What was built
- Cloud control plane design. One control plane to create, manage and govern resources across Azure, AWS, Google Cloud and Oracle, built on cloud SDKs and infrastructure as code, with security and cost controls built in.
- Topology automation. A low-code way to author each bank’s cloud topology and provision it from blueprints, in hours instead of weeks.
- Network segregation. Each bank’s environment kept apart on the network, across the hybrid links between the private core and the public clouds.
- Hybrid cloud blueprints. Where each workload belongs, which services to use, and how to deploy them, decided from data.
- A unified data model. One way to describe resources across clouds, so provisioning is standard and modular.
- The migration. Container-based services moved to each bank’s preferred hyperscaler.
- Working with teams. Architecture guidance, design reviews and planning across product, engineering, operations and risk.
Built with