AI Disrupts University Networks: Challenges and Innovations

The Challenges of AI for University Networks

The rise of artificial intelligence (AI) in higher education institutions is putting existing network infrastructures to the test. Students and staff demand seamless and uninterrupted wireless connectivity, which requires clever management of the wired infrastructures that connect campus networks. Dan DeBacker, product director at Extreme Networks, emphasizes that networks can remain idle for a time, then suddenly experience a surge in workload due to machine learning or other unforeseen uses. Whether this increase comes from machine learning, AI inference, or unexpected use cases, network administrators in higher education must demonstrate agility to keep everyone connected.

While many institutions prioritize Wi-Fi, all access points rely on a wired infrastructure. Therefore, handling the traffic generated by AI must be considered within the context of the entire network ecosystem. "Networks can remain idle for a while, then all of a sudden, everything spikes," says Dan DeBacker, product director for infrastructure at Extreme Networks.

Wireless Connectivity in the Age of AI

On university campuses, Wi-Fi is ubiquitous, and everyone desires fast and reliable access. However, the increased bandwidth demands due to AI present challenges. AI applications enable the customization of teaching to meet individual student needs and enhance scientific research. AI in operations-focused software helps reduce energy bills and refine course scheduling. According to a Cisco report titled "Wireless State 2026," deploying these applications requires high-performing wireless networks, which most campuses currently cannot provide.

Operational complexity, increased security threats, and a talent shortage limit institutions' ability to fully leverage modern networks. The report states: "Universities that face these three challenges together see significantly higher returns on investment." Georgetown University, for example, has modernized its wireless network by adopting Wi-Fi 7, suited for high-density and high-bandwidth environments. This upgrade helps reduce latency issues and improve the user experience. "Our infrastructure strategy is evolving towards an ‘experience-first’ model that sees the network as a dynamic platform for research and learning," says Douglas Little, CIO of Georgetown.

The Importance of Wired Infrastructure

Despite the emphasis on wireless, wired infrastructure remains essential for connecting various equipment and locations on campus. Managing these networks is becoming increasingly complex with the addition of numerous endpoints, such as laptops and IoT sensors. Integrating AI into these networks adds an extra layer of complexity. "No one really has a good idea of what this usage will look like," says DeBacker. Manufacturers like Extreme Networks are adapting their products to meet the new demands of AI-related workloads. Machine learning algorithms and sophisticated automation provide better visibility into network traffic and allow for real-time adjustments to enhance the user experience.

Looking ahead, DeBacker states that security will be one of the biggest networking challenges in the coming years. "With AI technology, we are going to see faster attacks," he says. AI-driven data breaches will spread much more quickly than their predecessors, so reacting when the lights turn red on a device or dashboard will be too late. "I would like to see when the box goes from green to yellow," he adds. AI-enhanced threat hunting will help defenders act decisively at that moment.

Tips for IT Leaders

Experts recommend several strategies to help IT leaders manage the impact of AI on university networks. For wireless networks, it is crucial to prioritize user experience and view the network as a foundational element for the university's critical systems. "Focus on providing a platform for ‘instant and sustainable’ connectivity that can support the next generation of AI tools and research," says Little. Dense environments, such as dormitories, are ideal testing grounds for assessing the network's capacity to handle high bandwidth demands. Little's team has collaborated with the Center for New Designs in Learning and Scholarship at Georgetown to ensure that network upgrades meet the real and evolving needs of faculty and staff.

Regarding wired infrastructure, it is important to anticipate constant changes and focus on flexibility. "We need to be able to adapt to bandwidth and traffic flows much, much faster than we ever have before," says DeBacker. The old model of upgrading the network every ten years is insufficient. DeBacker advises campus IT leaders and network administrators to focus on technologies that enable proactive and forward-looking oversight, rather than the traditional “firefighter” role of solving problems after they arise.

Finally, it is essential to educate staff about the risks and benefits of AI. "AI will permeate the network infrastructure," asserts DeBacker. "It's not an if, it's a when." There is no substitute for staying informed about technological changes and having a plan to address them, he adds. Ensure that people understand both the risks and the benefits of AI. And remind them that today’s students increasingly value bandwidth and will inquire about your current networking technology.