White Paper: Apstra AOS Architecture Overview

Introduction: Apstra AOS Is Designed to Solve the Hardest Data Center Problems

Building, deploying, operating, managing and troubleshooting a data center network can be difficult, expensive, and resource-intensive. According to a 2019 Gartner report, simplifying IT infrastructure is the number 1 strategic priority for businesses today. Apstra’s operating system (AOS) has been purpose-built to automate the full life cycle of the data center network.

In this paper, we discuss how the architecture of the Apstra operating system works to solve some of the most difficult problems faced by data centers today: composition, reliable change, extensibility and scalability.

AOS Architecture Goals: Composition, Reliable Change, Extensibility, Scalability

There are four main challenges facing data centers today:

  • Composition
  • Reliable Change
  • Extensibility
  • Scalability

Composition

Putting together a coherent, smooth working system with pieces of infrastructure from different vendors, and with different capabilities, can be incredibly difficult. Apstra’s operating system (AOS) creates a coherent whole - called a reference design - that enables an operator to deliver reliable and easy-to-consume services.

Reliable Change

Change is a constant in the data center. Change can come from an operator trying to add new services or expand capacity, or from the infrastructure in the form of failure conditions. Either way, operators have to find better ways to deal with change. According to a variety of studies in recent years.1 65-70% of network outages are caused by human configuration error while executing a planned change.

Extensibility

In order to encourage innovation and accommodate the inevitable rapid change in technology, data centers need to be able to add new capabilities and functionality easily and smoothly. As technology evolves, vendors innovate and come up with new features and capabilities that operators need to leverage in order to stay competitive. This evolution causes a design-time change resulting in a need for new reference design. In other words, your composition has to be extensible to allow for further innovation or to address new requirements.

Scalability

You want to be able to address all of this at scale. Data center networks need to be able to grow and accommodate innovation and increasing complexity. The primary goal of the Apstra operating system architecture is to address these issues directly, as we describe in the following pages.