Replatforming and refactoring

The infrastructure was migrated to AWS to achieve greater stability and flexibility. At the same time, the code had to be rewritten to decompose the MCZ monolith by containerizing each module separately. Now, each module scales independently based on demand, optimizing overall performance, reducing operating costs, and eliminating downtime.

Separation of layers using RDS and S3

The monolith was decomposed by separating the application layer from the persistence layer, with persistence now handled using Amazon RDS and S3. The goal was to isolate state from the application, eliminating the co-location constraints on the same server. Performance improves thanks to dedicated resources that scale independently, while costs are reduced by paying only for actual compute and storage usage. Furthermore, the new architecture ensures a fine-grained security model and enhanced resilience, relying on automatic backup strategies that are independent of the application’s lifecycle.

CDN Management

Amazon CloudFront provides a critical security layer for the architecture. As a CDN, it caches static content stored in S3 (such as images and PDFs), at global edge locations, minimizing load times for users. At the same time, it intercepts and blocks cyberattacks, such as DDoS attacks, directly at the perimeter, before they can impact the internal infrastructure. This protection not only secures the application but also optimizes overall cloud costs by preventing unnecessary or malicious traffic from overloading the core application containers and the database.

High Availability with Load Balancer and Autoscaling

The Load Balancer efficiently distributes traffic across containerized modules, handling failures and routing users only to healthy instances. Auto Scaling adjusts the infrastructure in real time, responding to traffic spikes and using only the resources required.

ECS Orchestration

ECS transforms MCZ’s isolated software components into a single, cohesive and resilient system, without manual intervention. The system decides where and when to run containers, optimizing AWS resource utilization. It ensures self-healing by instantly replacing a failed container, and coordinates updates with zero downtime, rolling out new features in the background while users continue browsing the site.

CI/CD Pipeline

With Continuous Integration (CI), every time the code is modified, the system automatically builds a new container image and runs rigorous tests to detect bugs and prevent regressions before release. Once the tests are passed, Continuous Deployment (CD) deploys the updated service to Amazon ECS, which updates it with zero downtime. This simplifies the release of changes, fixes, and new features on MCZ, eliminating the need for manual intervention or service interruptions.