OpenStack Core Concepts and Services

Understanding Core OpenStack Concepts and Services

OpenStack is a leading open‑source platform for building and managing public and private clouds. This course breaks down the fundamental ideas that appear in typical certification quizzes, helping you master the terminology, architecture, and key services such as Heat, Neutron, and the Placement service. By the end of the lesson you will be able to answer questions about cloud characteristics, the IaaS model, and the distinctive roles of the Undercloud and Overcloud in a TripleO deployment.

1. Cloud Computing Characteristics

One of the defining features of cloud computing is On‑Demand Self‑Service. This characteristic allows a user to provision resources—such as a virtual machine (VM)—through a web portal or API without any human interaction from the provider. The other common characteristics include:

• Broad Network Access: Services are reachable over the network via standard protocols.
• Measured Service: Resource usage is monitored, controlled, and reported, enabling pay‑as‑you‑go billing.
• Rapid Elasticity: Resources can be quickly scaled up or down to match demand.

When you see a quiz question that mentions starting a VM without contacting staff, remember that the answer is **On‑Demand Self‑Service**.

2. The IaaS Service Model

Infrastructure as a Service (IaaS) provides the foundational compute, storage, and networking resources. In this model, the cloud consumer is responsible for managing the operating system, middleware, and applications. The provider supplies the physical hardware, virtualization layer, and basic networking, while the consumer installs and configures the software stack that runs on the virtual machines.

Key take‑aways:

• The consumer does not manage physical hardware.
• The consumer does manage the OS, middleware (e.g., databases, web servers), and the applications they deploy.

3. TripleO Architecture: Undercloud vs. Overcloud

TripleO ("OpenStack On OpenStack") is a deployment model where OpenStack is used to install and manage another OpenStack cloud. It introduces two logical layers:

3.1 The Undercloud

The Undercloud is a small, dedicated OpenStack installation that acts as the installer and management plane. Its primary purpose is to deploy and manage the Overcloud. It runs services such as Ironic (bare‑metal provisioning) and Heat to orchestrate the creation of the production cloud.

3.2 The Overcloud

The Overcloud is the actual production cloud that end users interact with. It provides the compute, storage, and networking resources where tenants launch workloads. In other words, the Overcloud is the production environment, not a backup or a single‑node test setup.

4. Orchestration with Heat

Heat is OpenStack's orchestration engine, similar to AWS CloudFormation. It enables you to define the entire lifecycle of an application—instances, networks, security groups, and scaling policies—using a declarative template written in YAML.

Automatic scaling works through auto‑scaling groups defined in the Heat template. These groups monitor metrics (for example, CPU utilization) and trigger actions when thresholds are crossed, such as launching additional instances or terminating excess ones. This approach eliminates the need for manual flavor changes or manual cloning of images.

5. Networking Fundamentals: Neutron Virtual Routers

Neutron is the networking service that provides "network‑as‑a‑service" capabilities. One of its core components is the virtual router. Virtual routers route traffic between:

• Internal project (tenant) networks.
• External (provider) networks, enabling instances to reach the Internet.

While Neutron also offers DHCP agents, firewall services, and load balancers, the specific role of a virtual router is **routing traffic**, not storing block volumes or managing compute lifecycles.

6. Object Storage Explained

OpenStack Object Storage, known as Swift, stores data as discrete objects. Each object includes the data itself, metadata, and a unique identifier. This model differs from traditional block storage (attached to VMs) and file‑system storage (hierarchical directories). Object storage is ideal for unstructured data such as images, backups, and logs.

Key characteristics of object storage:

• Scalable to petabytes and beyond.
• Data is accessed via RESTful APIs (e.g., GET, PUT).
• Built‑in replication and eventual consistency.

7. Placement Service and Scheduling

The Placement service provides the scheduler with up‑to‑date inventories of available resources—CPU, memory, disk, and custom traits—across all compute hosts. When a new instance request arrives, the Nova scheduler queries Placement to find a host that satisfies the requested resources and any additional constraints (such as affinity rules).

Without Placement, the scheduler would lack a reliable view of resource availability, leading to over‑commitment or failed builds. Placement does not handle authentication, networking, or health monitoring; those responsibilities belong to Keystone, Neutron, and other services.

8. Recap of Core Concepts

Below is a quick reference that ties each quiz question to the underlying concept:

• On‑Demand Self‑Service – Users can launch VMs without staff assistance.
• IaaS Consumer Management – Manages OS, middleware, and applications.
• Undercloud – Deploys and manages the Overcloud.
• Overcloud – Production cloud for tenant workloads.
• Heat Auto‑Scaling – Uses template‑defined groups reacting to metric thresholds.
• Neutron Virtual Router – Routes traffic between internal and external networks.
• Object Storage – Stores data as objects with metadata and unique IDs.
• Placement Service – Supplies resource inventories for scheduling decisions.

9. Frequently Asked Questions (FAQ)

What is the difference between a virtual router and a DHCP agent?

A virtual router forwards packets between networks, while a DHCP agent assigns IP addresses to instances within a network. Both are part of Neutron, but they serve distinct purposes.

Can Heat modify an existing instance’s flavor?

Heat can replace an instance with a new one of a different flavor, but it does not directly resize a running instance. For live resizing, you would use the Nova API or CLI.

Is the Placement service optional?

Placement is required for accurate scheduling in modern OpenStack releases. Older deployments used the Nova scheduler’s internal inventory, but this approach is deprecated.

10. Further Learning Resources

To deepen your knowledge, explore the following official documentation and community guides:

• OpenStack Documentation Hub – Comprehensive reference for each service.
• TripleO Documentation – Detailed walkthrough of Undercloud and Overcloud operations.
• Heat Orchestration Guide – Templates, auto‑scaling, and resource types.
• Neutron Networking Guide – Virtual routers, networks, and security groups.
• Placement Service Overview – Resource providers, inventories, and traits.

By mastering these core concepts, you will be well‑prepared for certification exams, real‑world deployments, and advanced OpenStack engineering tasks.