Cloud-Native Network Provisioning
A self-service platform that turned spreadsheet-driven circuit turn-ups into audited, one-click workflows running on AWS and Kubernetes.
Project Overview
Every new enterprise circuit meant the same ritual: emails, spreadsheets, manual device logins, and a prayer that nobody fat-fingered a VLAN. The provisioning platform replaced it with declarative service definitions, peer-reviewed changes, and push-button deployment.
I designed the service model, wrote the Ansible and Terraform modules, and built the approval workflow that keeps auditors happy without slowing engineers down.
- •One-click audited turn-ups
- •Drift detection on schedule
- •Rollback to any prior state
The Challenge
Turn-ups took days, failed in creative ways, and left no trace of who changed what. The NOC inherited circuits with no documentation, and every audit season meant reconstructing history from chat logs.
- Manual device configuration with no version control, no review, and no reliable rollback path.
- Configuration drift between what was intended and what was live, discovered only during outages.
- Audit requirements demanding a full change trail that the old process simply could not produce.
Approach & Delivery
1 · Declarative service model
Defined every circuit as versioned code: endpoints, VLANs, QoS, and monitoring. Engineers propose changes as pull requests; nothing reaches hardware without review and automated validation.
2 · GitOps pipeline
CI renders intended configs, diffs them against live state, and posts the result on the PR. Merging to main deploys through staged environments with automatic rollback on health-check failure.
3 · Scheduled drift detection
Nightly reconciliation jobs compare live device state against the declared model and open tracked findings for anything that moved — no more mystery configuration.
4 · Self-service portal
A simple front end lets account teams request standard services from a catalog, with guardrails that keep requests inside tested, pre-approved templates.
Outcomes & Deliverables
- Standard turn-ups complete inside a single shift with a full audit trail attached to every change.
- Failed deployments roll back automatically, and drift findings give the NOC a prioritized fix list each morning.
- Reusable Terraform and Ansible modules plus runbooks, so new service types onboard in days instead of weeks.
Related Work
Want provisioning that audits itself?
I build automation that network teams actually trust — reviewed, tested, and reversible.
Book a consultation