Which OpenStack node runs the network agents?

In OpenStack, network agents are responsible for managing networking tasks such as DHCP, routing,

and firewall rules. These agents run on specific nodes within the OpenStack environment. Let’s

analyze each option:

A . block storage

Incorrect: Block storage nodes host the Cinder service, which provides persistent storage volumes for

virtual machines. They do not run network agents.

B . controller

Incorrect: Controller nodes host core services like Keystone (identity), Horizon (dashboard), and

Glance (image service). While some networking services (e.g., Neutron server) may reside on the

controller node, the actual network agents typically do not run here.

C . object storage

Incorrect: Object storage nodes host the Swift service, which provides scalable object storage. They

are unrelated to running network agents.

D . compute

Correct: Compute nodes run the Nova compute service, which manages virtual machine instances.

Additionally, compute nodes host network agents (e.g., L3 agent, DHCP agent, and metadata agent)

to handle networking tasks for VMs running on the node.

Why Compute Nodes?

Proximity to VMs: Network agents run on compute nodes to ensure efficient communication with

VMs hosted on the same node.

Decentralized Architecture: By distributing network agents across compute nodes, OpenStack

achieves scalability and fault tolerance.

JNCIA Cloud Reference:

The JNCIA-Cloud certification covers OpenStack architecture, including the roles of compute nodes

and network agents. Understanding where network agents run is essential for managing OpenStack

networking effectively.

For example, Juniper Contrail integrates with OpenStack Neutron to provide advanced networking

features, leveraging network agents on compute nodes for traffic management.

Reference:

OpenStack Neutron Documentation: Network Agents

Juniper JNCIA-Cloud Study Guide: OpenStack Networking