Santander Catalyst Platform
The Catalyst platform is Santanders internal, cloud‑native solution that abstracts infrastructure provisioning across more than 200 critical systems, enforcing architectural compliance while cutting deployment lead times from weeks to hours. Built on Amazon EKS and leveraging Crossplane, it delivers a unified, self‑service portal for developers, enabling rapid creation of new services and fostering a culture of automation.
Deep Technical Analysis
Catalyst centralizes control in a dedicated control plane cluster that orchestrates all provisioning workflows. By adopting infrastructure as code principles, the platform defines resources declaratively, ensuring repeatable, auditable deployments across multi‑cloud environments. The integration of Crossplane as a universal provisioner abstracts provider‑specific APIs, allowing developers to request resources through a consistent schema while the platform enforces policy, cost, and security guardrails.
Control Plane Architecture
The heart of Catalyst is an Amazon Elastic Kubernetes Service (EKS) cluster that hosts the orchestration engine, API gateway, and state management services. Kubernetes operators monitor resource requests, translate them into provider‑specific actions via Crossplane, and update the system of record. This design isolates developer interactions from underlying cloud complexities, providing a stable, versioned API surface.
Crossplane Integration
Within the EKS cluster, Crossplane extends Kubernetes' native API to represent external services such as databases, networking, and AI workloads. Custom Resource Definitions (CRDs) model each service type, and controllers reconcile the desired state with the actual cloud resources. This declarative approach reduces manual scripting, minimizes drift, and supports multi‑cloud portability.
Operational Impact
By automating the end‑to‑end provisioning pipeline, Catalyst slashed the provisioning cycle from 30 days to two days and reduced proof‑of‑concept setup from 90 days to one hour. Consolidating over 100 legacy pipelines into a single control plane lowered operational overhead, accelerated AI agent deployments, and cultivated a self‑service mindset among development teams, driving faster market response.