AWS & Cloud Engineering · Proof-of-Concept Lab
A documented cloud portfolio covering Linux and Windows compute, containers, private networking, infrastructure as code and managed data.
View the full project report
The Medi-Advice project brought together six hands-on infrastructure modules. I deployed web workloads, containerised a site, connected two VPCs over private routes, automated resources with CloudFormation and tested managed MySQL replication.
The goal was to move beyond isolated AWS console exercises. Each module needed a clear purpose, a working configuration and a final validation step that proved the service behaved as expected.
These were separate lab modules rather than one production environment. The CloudFront work was a documented delivery design, not a completed deployment, and is described that way here.
Core builds
Hosted the Medi-Advice site on Ubuntu EC2, transferred files over SFTP, then rebuilt delivery with Docker and Nginx.
Peered Manufacturing and Engineering VPCs, updated both route tables and validated private HTTP traffic in both directions.
Used CloudFormation Change Sets to add HTTP access, a 10 GB EBS volume, its attachment and an S3 bucket.
Created an RDS MySQL primary and read replica, queried replicated inventory data and confirmed that replica writes were blocked.
Validation evidence
Each module ended with a practical check. The evidence below keeps sensitive account details out of the public portfolio while preserving the technical result.
The Nginx container served the Medi-Advice landing page while Apache remained inactive on the host.
Routes and security-group rules allowed the Windows IIS and Ubuntu Apache servers to reach each other by private IP.
The final template managed the security group, EC2 instance, EBS volume, volume attachment and S3 bucket.
The replica returned the same ten inventory rows as the primary and rejected an attempted insert.
Production thinking
What I would strengthen next
The controlled lab used public access and broad rules in places to make validation possible. A production design would move databases into private subnets, restrict administration, use IAM roles and Secrets Manager, enable encryption and add monitoring, backups and Multi-AZ resilience where required.
Technology used
Next case study
Online Retail Sales Analysis →