Choosing Cloud‑Native Persistent Storage Solutions for Financial Container Platforms

Driven by digital transformation in finance, state‑owned and commercial banks are adopting containerization, facing the challenge of designing, building, and optimizing container cloud platforms that support agile, lightweight, and efficient development, testing, delivery, and operations.

In the cloud‑native era, selecting an appropriate persistent storage solution is critical; any commercial product supporting the Kubernetes CSI interface can be used. NFS is the most common protocol for simple persistence, while containers can leverage file, block, or object storage based on workload requirements.

Typical storage choices include:

File storage – suitable for small data and configuration files.

Block storage – ideal for high‑performance databases.

Object storage – for images, videos, and large files, requiring application support.

Kubernetes extends storage via CSI plugins, and mature block solutions such as TopoLVM (with Ceph) provide distributed storage capabilities for both compute‑storage integration and separation scenarios.

Database‑as‑a‑Service (DBaaS) platforms supply various databases (MySQL, TDSQL, TiDB, Oracle, MongoDB, Redis) and enable resource management, scheduling, and scaling, allowing connection counts to exceed 3000 per node.

Data consistency in high‑concurrency banking scenarios must address both business consistency (handled by microservices) and data consistency (handled by storage). StatefulSets with stable network IDs, ordered deployment, and scaling, combined with PersistentVolumeClaims, ensure pod‑level data persistence across restarts.

Beyond storage, the article recommends containerizing applications, adopting stateless designs, and moving toward micro‑service architectures to reduce database load and improve fault‑tolerance.