DataScale, a developer of digital solutions for the financial sector, has completed the key third stage of a project to create a high-load integration bus for Universal Bank. As part of the ‘Infrastructure Preparation’ stage, a fault-tolerant industrial platform based on Kubernetes was designed and implemented, which will become the foundation for all of the bank’s integration processes.
The goal of the project is to create a scalable, manageable, and high-performance bus capable of seamlessly connecting critical banking systems under peak loads.
Platform core: modern technology stack
The architecture implemented by DataScale is based on up-to-date solutions that have been proven in enterprise environments:
• Orchestration: Kubernetes version 1.31, providing automatic scaling and high availability.
• Network and security: Cilium plugin based on eBPF technology for effective network interaction and advanced security policies.
• Traffic: Ingress controllers based on NGINX and HAProxy, designed for loads of 20 to 60 thousand requests per second.
• Storage: Longhorn and NFS-CSI distributed systems for fault-tolerant data storage.
• Observability: a full monitoring stack based on Prometheus, Grafana, Loki, and Alertmanager for real-time monitoring of all components.
• Delivery: GitOps approach using Helm and ArgoCD, enabling secure and controlled continuous deployment (CI/CD) from GitLab Registry.
Architecture for high availability and growth
The DataScale team has worked out the infrastructure in detail, providing separate environments for development (DEV), testing (TEST) and production (PROD). The project includes:
• A clear network plan with no address space overlaps.
• Cluster sizing with dedicated nodes for control-plane, ingress, stateful applications, and general workloads. The architecture provides resources for seamless deployment of up to 900 pods in a production environment.
• Data infrastructure: a solution for storing service data and logs in MinIO S3 object storage, as well as detailed requirements for virtual machines and disk subsystems for each role in the cluster.
Business effects for the bank
The implemented Kubernetes platform gives Universal Bank not only technological advantages, but also direct business benefits:
1. On-demand scalability: the platform automatically adapts to load growth, ensuring stable operation during peak hours.
2. Fault tolerance: high bus availability minimises the risk of downtime and data loss.
3. Secure updates: the ability to release new adapters and changes without stopping services (zero-downtime deployment).
4. Product release speed: unified standards and automation accelerate development and integration processes.
5. Reduced operational risks and costs: centralised platform observability and manageability reduce the number of incidents and the total cost of ownership.
Roman Solovyov, CEO of DataScale, commented:
“The successful completion of the infrastructure foundation construction phase is an important strategic result for the entire project. We are providing the bank with more than just a set of technologies; we are providing a manageable, predictable, and fault-tolerant platform. This is the very digital foundation that allows new services to be implemented quickly and securely, minimising operational risks. Our expertise in building such systems for the financial sector has helped to create a solution that fully meets the stringent requirements for reliability and performance.”
About DataScale
DataScale specialises in developing and implementing digital solutions for banks and financial institutions. The company brings together experts with many years of experience in international IT projects and in-depth knowledge of modern technologies and global practices. DataScale creates comprehensive, scalable solutions focused on efficiency, security, and long-term customer development.
About AKB UNIVERSAL BANK
AKB UNIVERSAL BANK is a dynamically developing financial institution offering a wide range of services in the corporate and retail segments. The bank focuses on a high level of customer service, an individual approach and the active introduction of digital technologies to improve service quality.
