AI-Powered Campus Infrastructure Helps Budget-Strapped Universities Scale

Network Automation for Higher Education

Autonomous AI networks use machine learning to self-configure, self-optimize, self-protect and self-heal — all without manual human intervention. While fully autonomous networks are still evolving, many higher ed IT departments are already leveraging components such as self-healing tools, network automation and predictive analytics.

A good analogy for understanding autonomous networking is a self-driving electric car.

“We have a concept known as self-driving, which is the epitome of automation when it comes to networking,” explains Jeff Aaron, vice president of networking product and solution marketing at HPE. “There’s a journey to get there in terms of collecting data, and then you have to make meaningful insights out of that. You can then go towards driver-assisted networking, where you can give recommendations and a human goes to do the actual work. And ultimately you’re doing full self-driving, which sounds a lot like an autonomous vehicle like a Tesla. The same idea applies to networking.”

DISCOVER: How AI-native networking is already being deployed on campuses.

AI Network Automation in Practice

Among the many capabilities enabled by autonomous AI networks, self-healing offers one of the most practical entry points for colleges interested in boosting performance while easing the workload on IT staff. Self-healing uses AI to continuously monitor, detect and predict anomalies, then remediate them.

At Dartmouth College, HPE’s Marvis AI assistant is already putting this into practice. (Juniper Networks, maker of Marvis and the Mist AI platform, was acquired by HPE in July.)

“Our always-on Marvis AI proactively and continuously tests the network, looking for problems, and alerts IT staff before users do,” says Bryan Ward, lead network engineer at Dartmouth College. “The self-driving features of the Mist AI platform can, with permission, even take steps to remediate common issues without human intervention. We know that if the config works on one device, it will work on all of them.”

EXPLORE: How AI is already transforming campus networks.

Staffing Optimization: Reducing IT Operational Burden

As institutions weigh investments in AI-driven network solutions, one of the clearest advantages — especially for budget-conscious colleges — is the ability to dispense with manual interventions and free up valuable IT staff time.

“With a self-healing network, interventions are triggered automatically, with no manual intervention required,” says Kevin Jackson, senior solutions engineer at Fortra. “You free up your staff resources and take that workload off some of the day-to-day operators. Staffing can be reduced in many cases, and you can reallocate resources to better utilize them for strategic initiatives like planning for upgrades, focusing on business continuity and other big-picture initiatives.”

Automatic basic network troubleshooting enables institutions to shift Level 1 basic support to AI tools, decreasing the need for hands-on involvement from help desk staff and improving response times. Automated networking also brings efficiencies to the deployment process.

“We had a customer deploy about 200 access points in a couple of hours using templatized configurations,” says Aaron. “They even had students come in and do it. They plugged it in, scanned it, and it was up and running. They only needed to download the right configurations. The more automation you have, it just makes life easier for staff. They don’t want to spend their time having to go plug in access points and configure them. That’s a big job.”

Automation Increases Network Reliability

Offloading routine network configurations to AI not only improves efficiency but also enhances the reliability of the campus networks.

“Humans making config changes during business hours occasionally leads to service interruptions,” explains Ward. “Anyone who’s ever configured a network switch by hand knows the pain of accidentally removing all VLANs from an uplink port — it’s a rite of passage.”

Automated configuration not only prevents typos, but solutions like Mist also offer an automatic rollback feature that restores a device’s configuration to the last known working state if a config change takes it offline, according to Ward.

Network reliability gets an added boost from AI self-healing features, which help universities manage the network bandwidth more effectively across campus.

“Bandwidth is key. Campus bandwidth usage is something that can really get taxed,” says Jackson. “The network infrastructure that supports all the work on campus is routinely handling heavy workloads. Built-in automation can help throttle unauthorized bandwidth usage. So if network thresholds are reached, AI can simply reboot a router and dump some of those unauthorized connections, which allows you to recoup that bandwidth for authorized campus needs.”

DIG DEEPER: This AI Playbook provides broader higher ed AI strategies.

Challenges on the Road to a Self-Driving Network

While the benefits are clear, careful planning and a number of other factors must be considered before moving toward a fully autonomous networking environment.

“Institutions often face cost barriers when adopting new technology, whether they use a software-based solution or another hardware-based solution,” says Jackson. “Also, the timeline can be difficult because you want to avoid downtime during the school sessions to be able to implement and test your new setup. And it can be difficult to have the right expertise in place to deploy the solution.”

IT leaders must also assess what they are comfortable handing off to AI. Some network automations are easier to adopt than others.

“We let customers choose which actions go into self-driving mode,” says Aaron. “There are about 10 common actions, like rebooting a stuck port. AI can do that reliably. Why do you need a human to do it?”

Aaron lists a misconfigured VLAN or noncompliant software code among other examples of common self-driving options.

“The main actions are downloading the compliance software,” he added. “So these are some of the activities that can move more toward autonomous today, and we’ll be adding more.”

LEARN MORE: How a network assessment can uncover optimizing opportunities.

One of the biggest challenges of introducing autonomous features to the network, though, isn’t the tech — it can be the IT services culture itself.

“Many engineers, including myself, have built a career on networking a certain way,” Ward says. “We’ve toiled countless hours and cut our teeth doing things the ‘hard way,’ often with technology that hasn’t changed in over a decade. There is safety and comfort in keeping things unchanged.”

His advice? Start small.

“Changing how a team operates requires time, training and understanding,” he added. “Rolling out self-driving and automated networks is a long process but has immense value in the end. Start small and grow as you better understand and become more comfortable with the new technology.”