A move away from traditional three-tier infrastructure helped the University of Canberra reduce its data centre footprint by 70%, avoid expanding its data centre capacity, and eventually run around 95% of its on-premises workloads on Nutanix.
In this exclusive interview at Nutanix .NEXT 2026 in Chicago, Justin Mason, Associate Director, Technology Services at the University of Canberra, discusses the institution’s migration to Nutanix AHV, the operational benefits of infrastructure consolidation, and the technologies it is evaluating.
How did the University of Canberra’s Nutanix partnership begin?
It has been a long and successful partnership with Nutanix. Back in 2013, we were early adopters of the platform. We initially looked at it as a virtual desktop solution, but when we evaluated the platform more closely, we liked what it was doing and what it could offer. We reached out to Nutanix and told them we wanted to run some server workloads on it because we liked the hyperconverged concept. At the time, we were operating a traditional three-tier architecture. The hardware environment was quite disparate, which was not an efficient use of resources.
We conducted a small pilot on the Nutanix platform, and it went extremely well. In 2014, we moved production workloads onto the platform. Over the years, we started with Hyper-V on Nutanix before moving to VMware. We were also early adopters of Nutanix AHV (Acropolis Hypervisor), migrating from VMware to AHV around 2018 or 2019, well before Broadcom acquired VMware. Today, we operate separate clusters for different use cases, including a production cluster for production workloads, a research cluster for research activities, and a database cluster.
One of the most exciting projects we’ve undertaken recently involves providing self-service capabilities for researchers. Using Nutanix Cloud Manager Self-Service, researchers and some students can log into a web portal and deploy workloads themselves. This puts them in control of deploying Windows servers, Red Hat servers, and other workloads themselves
What challenges did your previous infrastructure create?
At the time, we were primarily running Hyper-V as our virtualisation platform, but we were buying separate physical servers and running Hyper-V on them. They were largely isolated, with one server supporting BMS (building management system) workloads, for example, and another supporting the student information system.
The challenge we faced was that we were running out of space in our data centres because we were constantly adding hardware. We could continue buying more racks, but there comes a point where you run out of space. You’re running into full space, and you need cooling, not more power.
We found that by migrating from the old three-tier architecture to Nutanix, our data centre footprint began to shrink. A few years ago, we calculated that by the time we had migrated 90% of our workloads onto Nutanix, our data centre footprint had been reduced by 70%. At one stage, we were building a larger data centre, but we realised we no longer needed it because we had been able to consolidate so much infrastructure.
Another benefit was the ease of management compared with the older three-tier architecture. Back then, many tasks were highly manual, whether taking systems offline for maintenance, applying patches, or performing firmware updates. Because we were running siloed servers, taking a system offline also meant taking the associated workload offline.
With Nutanix’s hyperconverged architecture, however, workloads are automatically moved between nodes during upgrades and maintenance. If one node is being patched or rebooted, the workload moves to another node, allowing maintenance to take place without downtime. This reduced the need for lengthy maintenance windows, and system administrators spent less time carrying out maintenance outside normal working hours.
Another advantage we discovered later was the ease of replacing or refreshing hardware within the Nutanix environment. Under the old set-up, deploying a new server meant manually migrating workloads across. With the Nutanix platform, when a node reaches the end of its lifecycle, typically after five to seven years, it can be safely removed from the cluster and replaced with a new node. The cluster then automatically rebalances and redistributes metadata without interruption or downtime.
What does your current infrastructure look like?
Infrastructure-wise, we’re currently running Nutanix on Supermicro servers. In terms of our on-premises infrastructure, we have two data centres: a main data centre and a micro data centre. One of the main reasons we were able to establish a micro data centre for disaster recovery was the consolidation provided by Nutanix. We didn’t need as much space.
In terms of what we run on Nutanix, being a university, it’s a mixture of different workloads. We run Oracle and Microsoft SQL workloads on the platform. Our HR system runs on it, as does our student information system. Our data warehouse, which is a pretty big workload, also runs on Nutanix. The primary server operating systems are Windows Server and Red Hat Enterprise Linux.
What percentage of your applications now run on Nutanix?
It’s probably around 95% now. Almost everything we provision on-premises these days goes onto the Nutanix platform. There are only a couple of exceptions: our CCTV system for security video and our back-up system. Neither runs on Nutanix.
That said, with Nutanix’s expanded support for external storage, our CCTV system could become a good candidate. It’s a storage-intensive workload that already runs on a dedicated storage platform. Being able to connect that to our existing Nutanix infrastructure and manage it there would reduce the number of separate systems we need to maintain. The expanded support for external storage was one of the announcements that caught our attention.
Our back-up system remains on a separate platform because back-up environments often have specific architectural requirements that need to be addressed independently.
How many applications do you run?
We have around 500 virtual machines across our production, test, and development environments. Production workloads account for around 100 to 150 applications. When you’re talking about virtual machines, most are running a specific application workload. In some cases, a system may consist of multiple virtual machines, such as an application server and a database server.
Many of our systems also have separate test and development environments, which means some workloads are duplicated. Across all environments, we would be looking at around 300 to 350 applications.
Of this number, 95% is running on Nutanix?
Yes, that’s correct.
How has the move from VMware to Nutanix AHV worked out?
That has been fortunate because some organisations have had to scramble to complete last-minute migrations, and you don’t want to be in that position. When we were running VMware on the Nutanix platform, it was rock solid and stable. There were no complaints, it was running well. However, when AHV came out, we saw it as another way to simplify the management of our infrastructure. Because it was built into the Nutanix platform, it offered a more integrated environment that would be easier to manage and update.
Like some organisations, we were a bit nervous at the time. I think the first thing we deployed on AHV was an Oracle RAC database cluster. Our DBAs were cautious because they were used to running databases on physical infrastructure. However, it proved to be a good test case. The Oracle RAC environment ran on AHV and performed well, and even the DBAs were happy with its performance and operation. That gave us confidence that if something as complex as Oracle RAC could run successfully on AHV, then migrating the rest of our workloads would be the right move. After that, everything else we deployed on the platform ran smoothly.
After that, we moved quickly to migrate everything from VMware to AHV, completing the process within six to 12 months. Most of the migration work was done manually, as tools such as Nutanix Move were not yet available. Even so, the process was straightforward and reliable. It mainly involved converting existing virtual machines into the Nutanix format and installing the required guest tools and drivers. Every workload we migrated came across successfully. Since then, we’ve continued to run AHV.
How much did the move to Nutanix save?
Before we migrated to Nutanix, like many organisations, we forecast our infrastructure spending over a three-year period based on the assumption that we would remain on a traditional three-tier architecture. When we compared those projections against our actual spending after moving to Nutanix, we found that we had saved around AU$1.2 million in infrastructure costs over three years.
We’re a small to medium-sized university, and our annual capital expenditure on infrastructure is typically around AU$2 million to AU$2.5 million. It’s not a gigantic amount, but for an organisation of our size, it was significant and provides some context for the scale of those savings.
Which recent Nutanix announcements stood out to you?
There were a couple of announcements that stood out to me. One was around VDI and AHV. We currently run some VDI (virtual desktop infrastructure) workloads on Dizzion, formerly Nutanix Frame, and we’ve also been discussing Azure Virtual Desktop internally because we use a lot of Microsoft products across the University. When I saw the announcement around running Azure Virtual Desktop on AHV, it seemed to align well with our existing infrastructure strategy.
The support for Nerdio also caught my attention. It could make migrating our existing on-premises VDI environment to Azure Virtual Desktop much easier.
Another announcement that stood out was Nutanix AI. During the keynote, I saw several applications in the demonstration that I think our Science and Technology faculty and researchers would be interested in using. The ability to deploy those applications in a more automated way through a central catalogue could make them easier to access and manage.
I was also interested in the landing zones announcement. Bringing the concept of public cloud landing zones into on-premises environments could be useful for organisations like ours that need to maintain separate segments within their infrastructure. I can see landing zones helping to segregate workloads for cybersecurity purposes.
