A Reproducible, Infrastructure-as-Code-Ready CI/CD System for a Smart Education Platform: Implementation and Local Evaluation

Smart education platforms need frequent, reliable releases, and institutions increasingly want to run them on more than one cloud. CI/CD reliability, infrastructure as code (IaC), cross-cloud DevOps, and cloud DevOps in education have been studied separately, but controlled end-to-end evaluations of an integrated workflow for an education application remain limited. This paper reports the locally evaluated first stage of such a study. We implemented a Smart Education Platform (React, Node.js/Express, MongoDB, Nginx), packaged it with Docker, and verified it with a five-job GitHub Actions workflow. Terraform configurations design AWS ECS Fargate and Azure Container Apps environments that share one commit-SHA image model, and an approval-gated, OpenID Connect-authenticated Azure release workflow adds a plan-scope policy, smoke tests, and rollback; neither environment has been deployed. A layered validation protocol was executed twice, a day apart, with identical results: 48 backend tests against MongoDB, 28 frontend tests, a 38-check real-browser journey, and offline IaC tests passed, and all 10 hosted CI runs succeeded. Static analysis found 40 failed Checkov policies, concentrated in encryption, and outdated base-image packages. In 12 k6 runs (795,308 requests, no failures), one Node.js process saturated a CPU core at about 480 requests/s, and Nginx cost 16–17% of throughput. In a controlled comparison against this baseline, four backend replicas raised step-load throughput 1.93-fold and lowered p99 latency. No cloud deployment or cloud benchmarking was performed, so cloud deployment time, cost, scalability, and portability remain future work.

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-07
DOI
https://doi.org/10.5281/zenodo.24029130
Primary Topic
Software System Performance and Reliability
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

A Reproducible, Infrastructure-as-Code-Ready CI/CD System for a Smart Education Platform: Implementation and Local Evaluation

Utsav Raj, Muskan Shrivastava
2 links
Zenodo (CERN European Organization for Nuclear Research)
Software System Performance and Reliability
preprint

A Reproducible, Infrastructure-as-Code-Ready CI/CD System for a Smart Education Platform: Implementation and Local Evaluation

Utsav Raj, Muskan Shrivastava
preprint en

Abstract

Smart education platforms need frequent, reliable releases, and institutions increasingly want to run them on more than one cloud. CI/CD reliability, infrastructure as code (IaC), cross-cloud DevOps, and cloud DevOps in education have been studied separately, but controlled end-to-end evaluations of an integrated workflow for an education application remain limited. This paper reports the locally evaluated first stage of such a study. We implemented a Smart Education Platform (React, Node.js/Express, MongoDB, Nginx), packaged it with Docker, and verified it with a five-job GitHub Actions workflow. Terraform configurations design AWS ECS Fargate and Azure Container Apps environments that share one commit-SHA image model, and an approval-gated, OpenID Connect-authenticated Azure release workflow adds a plan-scope policy, smoke tests, and rollback; neither environment has been deployed. A layered validation protocol was executed twice, a day apart, with identical results: 48 backend tests against MongoDB, 28 frontend tests, a 38-check real-browser journey, and offline IaC tests passed, and all 10 hosted CI runs succeeded. Static analysis found 40 failed Checkov policies, concentrated in encryption, and outdated base-image packages. In 12 k6 runs (795,308 requests, no failures), one Node.js process saturated a CPU core at about 480 requests/s, and Nginx cost 16–17% of throughput. In a controlled comparison against this baseline, four backend replicas raised step-load throughput 1.93-fold and lowered p99 latency. No cloud deployment or cloud benchmarking was performed, so cloud deployment time, cost, scalability, and portability remain future work.

Zenodo (CERN European Organization for Nuclear Research)
Alliance University (IN)
Quality education, Industry, innovation and infrastructure
Software System Performance and Reliability
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon