Notre Dame swaps 955 campus switches in about five months to end building-to-building drift

A user walks out of one Notre Dame building and into another, and something that worked a minute ago stops working ^[8]. For the campus IT team that was a standing complaint. "We would routinely receive service tickets about things working for users in one building, and then go to another building, where it was not working," John Buysse, senior director of network and telephony services, told SiliconANGLE ^[8].

The team's own policy produced that drift. "Our previous approach, which led to a wide variety of hardware, was just yearly upgrades," Buysse said. "That took a tremendous amount of our time." ^[9] The column calls it a sensible-sounding policy, not neglect ^[18]. By the end the fleet spanned roughly 28 models and more than 100 operating system versions, all of them Cisco ^[3]. That averages out to more than three and a half software versions per hardware model ^[3]. The goal now is that "regardless of what building a user may end up in, they have a consistent experience," Buysse said ^[10].

The column hangs the project on AI. Its argument is that becoming "AI-ready" is rarely about a new model or agent and more often about ripping out a decade of infrastructure debt ^[16]. Its occasion was a multiyear Cisco partnership with Notre Dame Athletics ^[5]. The staff quoted gave narrower reasons. "As we upgraded the campus to 6-gigahertz-capable, we needed improved back-end bandwidth," Buysse said ^[11]. Power was the second limit. "We were heavily underutilizing switches because we couldn't fully populate them without blowing the PoE budget," he said, with cameras and phones drawing on that budget alongside access points ^[12]. Campus bandwidth use more than doubled between fiscal 2021 and fiscal 2026, largely from video ^[6].

Tracy Weber, the chief information officer, puts the bar with the students. "The washers and dryers in the residence halls need to connect to the app that tells them how many minutes are left," she said ^[14]. She measures success by the student newspaper. No complaints in The Observer "is a good sign because it does hit the news when they are not happy," she said ^[15]. The column does not report ticket counts from before or after the cutover.

The pace is the part an operator can plan against. The first building cut over during spring break in March ^[4]. Over about five months, that comes to roughly 190 switches a month, or about 5.6 per building on average ^[1][2]. The tooling named is Catalyst Center with zero-touch provisioning ^[2]. According to the column, the campus never went offline during the swap ^[1]. Weber described the constraint plainly: "This place never shuts down." ^[13]

Two counts from a team's own records sort the decision. One is variance: distinct hardware models and OS versions per site. The other is the share of tickets describing a service that works at one site and fails at another. High variance with location-dependent tickets was Notre Dame's position, and it is where a single fleet-wide refresh has the clearest case. High variance with tickets that do not track location means the fleet costs staff hours but not users, and the swap can wait for the lifecycle date. Low variance with location-dependent tickets points at configuration or wireless, and new switches would not fix it. Low on both leaves nothing to buy. The column's advice is to run the numbers on a consolidated refresh, labor included, before defaulting to another year of piecemeal upgrades ^[17]. Notre Dame planned its new network around a 10-year lifecycle, with a 2030 goal of 100-gig connections to end users where they are needed ^[7].