Evolution of NetBank's Cloud-Native Architecture

Since its inception, NetBank has built its core systems on cloud infrastructure, initially using a classic virtual‑machine based deployment model. As business volume grew, the two‑site three‑center architecture could no longer meet scalability and disaster‑recovery requirements, prompting a shift to a multi‑active cloud‑unit architecture.

With the rapid maturation of cloud‑native technologies, NetBank is now constructing a financial‑grade cloud‑native distributed architecture to provide a secure, stable, efficient, and agile infrastructure for the new era.

The cloud‑native stack includes containers, Kubernetes for unified resource scheduling, service mesh for transparent communication, immutable infrastructure, and declarative APIs, enabling fault‑tolerant, easily managed, and observable systems.

Key challenges of the upgrade involve large‑scale infrastructure replacement, ensuring smooth coexistence between legacy and cloud‑native components, transitioning from package‑based delivery to container‑image delivery, and building new operational and security capabilities.

Upgrade goals focus on deploying cloud‑native infrastructure, achieving seamless migration, enhancing development efficiency, resource utilization, security hardening, and establishing an automated, sustainable delivery pipeline.

Practical implementations include a hybrid‑cloud elastic architecture that decouples applications from any single IaaS provider, immutable image‑based deployments that simplify multi‑environment rollouts, and a service mesh that offloads RPC, messaging, and database access to sidecars, improving transparency and multi‑language support.

Security is reinforced through a trustworthy cloud‑native design, with service‑mesh‑based access control providing deep defense for core services while maintaining development speed.

To improve resource efficiency, NetBank mixes online and offline workloads on shared clusters, applying dynamic scheduling and capacity‑calculation models to achieve high utilization and cost reduction.

Looking ahead, the bank expects cloud‑native abstractions to further separate infrastructure concerns from business logic, enabling automated operations, serverless innovations, and a middle‑platform architecture that supports rapid iteration of complex, distributed financial services.