Cloud Computing Lab

Code and lab records from the Cloud Computing Lab at REVA University, Bengaluru (B.Tech, sixth semester, 2020) — the same semester as Data Mining. A cloud course is part algorithms and part systems, so this is a mix: Python simulations for the concepts (scheduling, load balancing, MapReduce, a small cloud simulator), a bit of Java for the Hadoop side, real Docker / Kubernetes / Terraform artifacts, shell scripts for the AWS/GCP CLIs, and markdown lab records for the hands-on VM and platform experiments.

Nothing here touches a real cloud account — the AWS/GCP scripts are the commands you'd run (with placeholders), and everything else runs locally or is simulated in memory.

Structure

Folder	What's in it	Mostly
01-introduction/	Service/deployment models, cost calculator	Python
02-virtualization/	Hypervisors, VM placement, migration, overcommit	Python
03-distributed-computing/	CAP, Lamport & vector clocks, consistent hashing, quorum, 2PC	Python
04-mapreduce-hadoop/	Word count, inverted index, matrix multiply (Python + Java)	Python/Java
05-scheduling/	FCFS, SJF, round robin, priority, min-min, max-min	Python
06-load-balancing/	Round robin, weighted, least-connections, IP hash, throttled	Python
07-cloud-simulation/	A small from-scratch CloudSim (datacenter, VMs, cloudlets)	Python
08-containers-orchestration/	Dockerfile, compose, k8s manifests, a container scheduler	Docker/K8s/Py
09-cloud-storage-security/	Object store, replication, encryption, IAM, erasure coding	Python
10-cloud-platforms/	AWS/GCP CLI scripts, Terraform	Shell/HCL
lab/	The 14 lab items — Python sims + platform lab records	Python/Markdown

Each topic folder has a NOTES.md with the concepts and the algorithms.

Running

The Python simulations are standard-library only — no pip install needed:

python3 05-scheduling/05_min_min.py            # min-min task-to-VM mapping + makespan
python3 06-load-balancing/01_round_robin.py    # requests cycle across the servers
python3 07-cloud-simulation/04_cloud_sim.py    # datacenter + broker, prints makespan and cost

The Java word count is plain Java (no Hadoop JARs), so it compiles and runs standalone:

cd 04-mapreduce-hadoop && javac WordCount.java && java WordCount

The container app is a tiny stdlib web server that the Dockerfile packages:

cd 08-containers-orchestration
python3 app.py            # serves on :8080  (or docker build -t cc-web . && docker run -p 8080:8080 cc-web)

The infrastructure files

08-containers-orchestration/ and 10-cloud-platforms/ hold the real artifacts — a Dockerfile, a docker-compose.yml (web + redis), Kubernetes Deployment/Service manifests, and Terraform (main.tf, variables.tf) for an EC2 instance and an S3 bucket. The shell scripts in 10-cloud-platforms/ are AWS-CLI and gcloud command sequences with placeholder values (ami-xxxxx, <your-key>, my-bucket) — they're valid bash but need real credentials to actually run, so they're kept as examples.

Lab

lab/ is a mix, which is how the lab actually ran: six Python simulations (CloudSim scheduling, MapReduce word count, load balancing, VM allocation, consistent hashing, replication) and eight markdown lab records for the hands-on parts — creating a VM in VirtualBox, building and running a Docker container, deploying to Minikube, launching EC2 and S3, Google App Engine, and OpenStack with DevStack. The records have the real commands, dated across the semester.

A note on the code

Sixth-semester lab code — short names (vm, host, lb, sched, mk), the simulation written out plainly rather than cleverly, the YAML and Dockerfiles hand-written the way you'd actually write them. It's meant to be read and run, and it is the code as submitted.