The transcendence of cloud computing and Software-as-a-Service (SaaS) has somewhat obscured a quieter revolution taking place in traditional data centers as new automation tools whittle away at the complex and labor-intensive (read "expensive") tasks of building, deploying, operating, managing, protecting, and verifying the reliability of the center and its network.
The typical data center is built with pieces of infrastructure from many different vendors, each with different configurations and capabilities, and new uses and functions are constantly being added in pursuit of digital transformation. IT departments find themselves with the Sisyphean task of managing millions of connected devices in near-real time. Operators find themselves dealing with dozens of applications, each with its own requirements and functions.
Translating a company's policies and goals-its "intentions"-into reality can often involve an army highly-paid engineers going through and deploying thousands of new lines of code throughout the infrastructure.
The hottest current solution is something called intent-based networking (IBN), according to Zack Zilakakis, a former Gartner analyst and spokesman for Apstra, the Menlo Park, CA-based startup that pioneered the IBN approach when it came out of stealth in 2016.
Zilakakis explained how Apstra's IBN approach works in a telephone interview:
Simply put, intent-based infrastructure allows customers to describe "what" they want their network to do, and Apstra translates this to the "how" through automation and validation. The solution leverages machine learning to adapt and become more aware over time to optimize the experience, based on the defined policy and intent. In essence, Apstra introduces simplicity and reliability through automation and validation.
A real differentiator is that our solution is tailored to architects and operators covering the full lifecycle from Day 0 through Day 2. The architects use it to automate and validate the design, build, and deployment of the network infrastructure. The operators use it to reliably make changes, ensuring the network operates as intended through continuous automation and validation. Apstra removes the barrier between architecting and operating the network.
In the old days, two or three years ago, intent-based networking may have sounded like the usual marketing hype but the concept got a major boost in 2017 when Cisco CEO Chuck Robbins said it would "redefine the network for the next 30 years" and began rebuilding the networking giant around the concept.
The IBN movement actually began in 2014 when Apstra's three founders, Mansour Karam, David Cheriton, and Sasha Ratkovic, came together to build the company's core Apstra Operating System (AOS) from the ground up to support IBN. Said Zilakakis:
Once AOS is up and running it keeps a real-time repository of configuration, telemetry and validation information to constantly ensure the network is doing what the customer wants it to do. It also includes automation features that provide consistent network and security policies for workloads across physical and virtual infrastructures as well as intent-based analytics to perform regular network checks to safeguard configurations. And, AOS is hardware agnostic and integrated to work with products from Cisco, Arista, Dell, Juniper, Microsoft and Nvidia/Cumulus.
In August 2020, Apstra added integration of Juniper Networks and SONiC to its operating system (AOS 3.3) and bundled in a number of new functions. Among the Juniper enhancements, AOS includes support for Ethernet VPN (EVPN) which ensures that EVPNs are built properly on all devices and validates the EVPN routing topology across the entire network. It also lets Juniper customers ensure Juniper devices can't be added to a network until certified by the customer's security policies.
AOS 3.3 supports zero-touch provisioning that lets new devices fresh out of their boxes be automatically updated to the proper Juniper Junos version and validated without any operator input.
The latest update also provides support for the Microsoft-pioneered Software for Open Networking in the Cloud (SONiC) software that runs on Linux aimed at open network switch operations and management. Said Zilakakis:
SONiC is now getting validation in the industry. It's moving past hyperscale data centers into large enterprises and even toward a lot of smaller and mid-sized companies and organization. It's not just the Fortune 500 or the Fortune 100 anymore but organizations of all sizes. So we are really completely fully committed to the SONiC revolution that's going on in open networking. We're really, really excited about what we're doing there because we're able to bring this abtraction layer and this operational model to the enterprise, and give them more flexibility in their digital transformation initiatives.
The bottom line, the company claims, is that networks are twice as agile, 70% faster and 83% cheaper to run. Said Zilakakis:
IBN is part of a trend toward intelligent automation that brings together network architecture and operations at Day 0 in a common cause to abstract complexities and make both infrastructure and operations simpler, reliable, more flexible, and agile.
Global Market Insights predicts that the IBN software segment will grow at a CAGR of above 25% from 2020 to 2026 and will cross US$4 billion by then.
Intent-based networking is another one of those clever technological innovations that have clear benefits for enterprises in terms of efficiency, productivity, and cost-savings which, in theory, also have pass-along benefits to customers. However, the cost-savings comes in the form of fewer and less well-paying jobs for both IT and operations. Big tech likes to say that smarter software means less grueling and repetitive work that allows people to focus on the really important things. In many cases, the grueling and repetitive work is the important thing.
The McKinsey Global Institute 2019 Future of Work report found that some 50 percent of current work activities are technically automatable. That number seems only likely to grow as tech firms leverage machine learning and new A.I. platforms to build more ever more sophisticated tools.
Meanwhile, the Robotic Process Automation revolution marches on. Gartner says the RPA software market grew 62.9% in 2019 to $1.4 billion and held its position as the fastest-growing segment in the enterprise software market for a second year in a row.
Tech cheerleaders like to argue that change is both inevitable and good. Inevitable? Sure. But, the increasingly sordid history of social media suggests that it isn't always good.