Learn OpenStack - History, Concepts & Main Architecture

#Introduction

In episode 0, you were already standing at the gate: DevStack is running and openstack service list shows the lineup of services. Now it's time to understand why OpenStack exists, what it does, and how it's structured. This is the conceptual foundation that separates a knowledgeable operator from a mere copy-paster of commands.

Episode 1 covers three things: the history and background of OpenStack, the definition of OpenStack as an IaaS platform, and the main architecture with four types of nodes. At the end of the episode, you'll also see OpenStack's position compared to other platforms such as VMware vSphere, CloudStack, and Proxmox.

#History & Background of OpenStack

#Birth in 2010

OpenStack was born in 2010 as a joint project of Rackspace and NASA. Rackspace, a hosting company, and NASA, which needed flexible computing infrastructure, combined two codebases: Swift (object storage) and Nova (compute). Both were released as open-source and became the foundation of the first OpenStack.

Since then, OpenStack has been governed by the Open Infrastructure Foundation (OpenInfra Foundation) — a non-profit organization that also hosts projects such as Kubernetes sub-projects and Airship. Its community consists of hundreds of vendors and thousands of contributors from around the world.

#Growth and Name-Based Releases

OpenStack releases versions with city names, following alphabetical order:

2010 Austin
2011 Diablo
2012 Folsom
2014 Icehouse
2016 Newton
2018 Rocky
2020 Ussuri
2022 Zed
2024 Caracal
2026 - and so on

Each release brings stability improvements and new features. This means you need to pay attention to the version when deploying to production, because APIs and service behavior can change.

#What is OpenStack?

#Private Cloud Infrastructure-as-a-Service

OpenStack is an open-source IaaS (Infrastructure-as-a-Service) platform that centrally manages Compute, Networking, and Storage through APIs and a Dashboard. Unlike managed platforms such as AWS, you deploy OpenStack yourself on your own hardware — that's what's called a private cloud.

With OpenStack, you can provision internal cloud resources:

• Self-service: users request their own instances, networks, and volumes through the dashboard or API without admin intervention.
• Multi-tenancy: resources are isolated across projects with quotas and RBAC.
• Automation: all operations can be performed via the API, paving the way for orchestration and IaC.

openstack endpoint list --interface public

openstack endpoint list --interface public shows all API endpoints — proof that OpenStack is truly API-first. The Horizon dashboard is merely a client of these APIs.

#Key Concepts You Must Understand

Some terms that will keep appearing throughout the series:

• Project (Tenant): the resource isolation unit, like a cloud account.
• Instance: a VM running on the hypervisor.
• Image: an OS template for creating instances.
• Flavor: instance hardware specifications (vCPU, RAM, disk).
• Network: the virtual network that connects instances.

#OpenStack Main Architecture

#Four Node Roles

Conceptually, an OpenStack cluster is divided into four roles:

#Control Plane

The Control Plane is the brain of the cluster. It runs all the API Services, the Scheduler that picks the destination node for an instance, the MariaDB/Galera database that stores state, and the RabbitMQ message queue that sends instructions between services.

user → API (Keystone/Nova) → Scheduler → Message Queue → Compute Node
              ↓
        Database (state)

If the Control Plane goes down, the APIs become unreachable, even though already-running instances stay alive. This is why production uses more than one controller node (HA).

#Compute, Network, and Storage Nodes

Compute Nodes are the workers: they run VM instances using the KVM/QEMU hypervisor. Network Nodes manage SDN (Software-Defined Networking) — routing, DHCP, and NAT virtually. Storage Nodes provide storage: block for volumes, object for unstructured data, and shared filesystems for shared access.

In real production deployments, these roles can be combined or split depending on scale — we'll dissect this topology in episode 14.

#Comparison with Other Platforms

#OpenStack's Position in the Ecosystem

OpenStack isn't alone. Three platforms are frequently compared against it:

VMware vSphere is the dominant player in traditional data centers, with a mature hypervisor and rich features, but it's proprietary and has high licensing costs. CloudStack offers an IaaS that's simpler to operate. Proxmox is very popular at small-to-medium scales because of its ease of use.

OpenStack wins at these points: vendor-neutral, complete APIs, enterprise multi-tenancy, and an open ecosystem. Its downside is that the learning curve and operational complexity are the highest of them all. This is why this series matters — you learn to manage that complexity step by step.

#Summary

Episode 1 places OpenStack in context: the 2010 birth story by Rackspace and NASA under the OpenInfra Foundation, its definition as an API-first open-source IaaS platform, the four-node-role architecture (Control Plane, Compute, Network, Storage), and its position against vSphere, CloudStack, and Proxmox.

Key takeaways:

• OpenStack was born in 2010 from Rackspace and NASA and is now under the OpenInfra Foundation.
• OpenStack is open-source IaaS for private clouds, operated via APIs and the dashboard.
• The Control Plane contains APIs, the scheduler, MariaDB/Galera, and RabbitMQ.
• The Compute Node runs instances with KVM/QEMU.
• OpenStack is chosen because it's open, API-first, multi-tenant, and enterprise-ready.

In episode 2, we'll cover OpenStack core services and inter-component communication — Keystone, Glance, Nova, Neutron, Cinder, and Horizon, and how they talk to each other through the RabbitMQ message queue, the MariaDB/Galera database, and RESTful APIs. Make sure openstack service list still works well in your lab.