Sber

How we built a scalable microservice architecture for secure biometric authentication

Sber Biometric Authorization System

For Sber, we developed an enterprise system for biometric authentication and identity verification that processes millions of sessions daily — with maximum security, stable availability, and full auditability. The platform combines modern microservices, encrypted communication, and intelligent load balancing to enable real-time biometric logins across web and mobile platforms.

Challenge

Sber needed a new authentication layer that:

  • Reliably processes millions of biometric requests daily
  • Is secure, scalable, and fault-tolerant
  • Supports full logging & compliance
  • Can be integrated into a containerized infrastructure
  • Seamlessly integrates with existing identity providers

In short: a modern, modular platform for critical security processes.

Our Approach

1 — Microservice-Based Architecture

We developed independent services for:

  • biometric validation
  • session and token management
  • document and ID verification
  • event and audit logging

Communication between services occurs asynchronously via message queues.

2 — Containerization & Orchestration

All services were implemented using Docker and operated in Kubernetes:

  • automatic horizontal scaling
  • high fault tolerance (self-healing)
  • rolling deployments without downtime

3 — Security & Compliance

The system was developed with bank-grade security standards:

  • encrypted service-to-service communication
  • multi-factor validation
  • adaptive throttling mechanisms
  • full audit traceability

4 — Integration & Extensibility

The platform was designed to flexibly integrate into:

  • existing identity providers
  • internal banking systems
  • mobile and web apps

Results

  • 2.5+ million biometric sessions per day
  • Downtime reduced by 60% thanks to containerized architecture
  • Full auditability for internal compliance processes
  • Seamless integration with external and internal identity providers
  • Stable performance under high load

Tech Stack

Backend: Java 17 · Spring

Database: Oracle

Infrastructure: Docker · Kubernetes

Messaging: asynchronous queues

Duration: 18 months

Team: 5 engineers

Why It Matters

The architecture developed in this project sets standards for modern security-critical systems. The same principles — modularity, security, scalability, and clear responsibilities — now flow into our startup and enterprise projects, where reliability and data protection are crucial from the start.

Case Studies

Sber: Scalable Microservice Architecture for Biometric Authentication | Enterprise Banking | H-Studio