Cloud Infrastructure Administration
This was a ten week course (IT227) on administering cloud infrastructure, working mostly across AWS and Azure. It started from the basics of what the cloud actually is and built toward a full migration plan for a multi-tier application, so the arc went from concepts to something close to a real design decision.
What the course covered
The early weeks were foundational: on-demand shared resources, the IaaS, PaaS, and SaaS split, and the deployment models. The hybrid model got real attention, which lined up with my own work keeping sensitive data on-premises for compliance while running less restricted workloads in a government cloud.
From there it got more hands-on with provider services. I spent time on AWS building blocks like ECS, S3, RDS, and CloudWatch, then did a direct AWS versus Azure comparison including how the two map their terminology, an EC2 instance versus an Azure VM, a VPC versus a VNet, IAM versus Entra ID. Security came next: encryption in transit versus at rest, TLS, and why the transmission path is often the weakest link even when the storage itself is locked down.
The back half moved into operations. We covered vertical versus horizontal scaling and when each one makes sense, data storage choices across SQL Server, MongoDB, and Redis, and how those fit different data patterns rather than one tool doing everything. Then logging, monitoring, and diagnosing: CloudWatch, X-Ray, and third-party APM tools like Splunk and Datadog for tracing requests and spotting bottlenecks. It closed with SLAs and the shared responsibility model, and a final migration brief that pulled the whole stack together.
What I took from it
The most useful part was learning to justify a platform choice instead of just naming services. The migration work forced me to weigh onboarding effort, operational overhead for a small team, and monitoring strategy against each other, not just pick whatever was technically possible.
The monitoring and observability material has been the most directly applicable to my day job. Knowing how to trace a slow request through the tiers and read database load is worth more than memorizing any single service name.