Bank of Anthos

Bank of Anthos was conceived to bridge a critical gap in African financial inclusion. While digital finance solutions are emerging, millions across the continent, from market traders to established professionals, remain underserved. Our mission is to empower every African to fully participate in the benefits of the digital economy. Through features like transaction-history-based credit scoring for swift loan access and instant bank statement generation for proof of financial standing, Bank of Anthos directly addresses key barriers to financial progress, fostering a future where digital finance truly serves all.

Methodology

The Bank of Anthos project utilized a cloud-native, microservices-based architecture deployed on Google Kubernetes Engine (GKE). The development and deployment followed a robust CI/CD pipeline with automated builds and deployments triggered by code changes in individual microservices. Here are the highlights:

Microservices Decomposition: The application was broken down into distinct services (e.g., frontend, userservice, ledgerwriter, balancereader) each with specific responsibilities and often written in different languages (Python, Java).

Managed Services for Infrastructure: The project heavily leveraged Google Cloud Platform (GCP) managed services like Cloud SQL (PostgreSQL) for databases, Cloud Build for CI/CD, Artifact Registry for container image storage, Cloud Monitoring, and Cloud Logging for observability. This approach reduced operational overhead and leveraged GCP's inherent scalability and reliability.

Automated CI/CD: Each microservice had its own isolated Cloud Build configuration, ensuring that changes to one service didn't inadvertently affect others during the build and deployment process. kubectl rollouts were used for seamless updates.

Security by Design: HTTPS enforcement, JWT-based authentication, Workload Identity for least privilege access to Cloud SQL, and automatic TLS certificate management were integral to the security posture.

Observability: Custom Cloud Monitoring dashboards and structured Cloud Logging were implemented to provide real-time insights into service health, performance, and error rates, enabling proactive issue detection and resolution.

Load Testing: The inclusion of a loadgenerator service (Python/Locust) demonstrates a methodology for simulating user behavior and testing the system's performance under load.

Challenges

Microservices Complexity: Managing multiple independent microservices, their interdependencies, and communication protocols (e.g., REST endpoints, JWT validation across services) can introduce significant complexity compared to a monolithic application.

Database Management and Migration: The migration from on-cluster PostgreSQL to Cloud SQL, while beneficial for high availability, would have required careful planning and execution to ensure data integrity and minimal downtime. Managing two separate Cloud SQL databases (accounts-db and ledger-db) adds a layer of data management complexity.

Results

the Bank of Anthos project demonstrates a successful application of modern cloud architecture principles, resulting in a flexible, secure, and observable system capable of simulating complex banking operations.