The Apstra Operating System

Enterprises continue to struggle with the network as the bottleneck in data center operations.

The Apstra Operating System (AOS) is a vendor-agnostic distributed operating system for the data center network that enables business agility, dramatically scales operational efficiency, and reduces downtime.

How it Works

The Apstra Operating System (AOS) automates the entire lifecycle of network infrastructure and services.  Starting with customer-declared statements of intent, AOS automates the various processes involved in designing and deploying network configurations, and ensuring that the network is always “ready for business” through real-time telemetry and analytics that provide continuous validation.

AOS is the first product that integrates configuration and state for automating all aspects of Day 0, Day 1, and Day 2+ data center operations as well as continuous validation of business readiness.

AOS 1.1 Features 


Fabric connectivity

  • Server and rack-based design intent
  • BGP L3 CLOS fabric
  • L3 (routing on the host) server attachment
  • L2 server attachment with MLAG/LAG
  • DHCP relay

Virtual Networks

Device OS

Cumulus Linux and CVX
Arista EOS and vEOS
Cisco NX-OS
Ubuntu servers

AOS ETC (Extensibility Tools for the Community)

Zero Touch Provisioning (ZTP) Server
Demo Tools
Template Catalog
3rd Party Tool Integration
3rd Party Big-Data Platform Integration
Legacy Devices Integration


LLDP, BGP, Config deviation
Interface counters
Routing table verification
Intent-based anomaly detection
Streaming via Google protocol buffers
Servers and devices health (cpu, mem, disk)


Single User Authentication
Device Lifecycle Management
Resource Management
Headless Operation
Scalability up to 1600 devices
System configlets
Interactive topology visualization

Maintenance workflows

Scaleout Maintenance
Replacement Maintenance
Decommission Maintenance
Addition and deletion of virtual networks

“Apstra has developed the best automation platform for building and operating networks…To put it plainly: it works. It’s easy to install and use, provides great network visibility, works on equipment from many vendors, and improves security. We use AOS at Awnix, and we’re incorporating it into the designs for our Engineered Systems for OpenStack with SDN.”

Rick Kundiger, Awnix CEO   


Who will use AOS?

Network engineers and architects, network operators

Who else should care about AOS?

CIOs who want to build private clouds with networks that are simpler to manage and operate and want a choice of hardware.

How does AOS accomplish this?

AOS enables infrastructure teams to design, build, deploy, and operate infrastructure using an integrated solution purpose built to automate network services.

The user starts with specifying business intent. AOS then designs vendor-agnostic network infrastructures that satisfy intent, builds specific blueprints that pull in the appropriate resources, and allows infrastructure teams to deploy blueprints with a touch of a button.

AOS then collects rich telemetry at a high rate to validate in real time that the network is continuously ready for business.

What alternatives exist in the market?

There are currently some home-grown tools that attempt to automate specific use cases, separate configuration and state, and require a lot of integration work that doesn’t operationally scale.

There also some server infrastructure tools on that market that do not integrate configuration and telemetry in a continuous validation loop and that do not start with business-level intent.

What are AOS’s key differentiators?

AOS is intent-driven and integrates configuration and state, collects real-time telemetry and provides real-time continuous validation, and is vendor-agnostic, allowing the customer to leverage a combination of hardware to best suit their needs.