Infrastructure Services is responsible for managing a wide range of services and underpinning IT infrastructure used daily by staff and students. Everything from eduroam, Single Sign-On (and multi-factor authentication), Nexus 365 (including Teams), the University’s network connection to the Internet (and every building’s connection to the University network), through to research computing, data backup, telephones, and the Apple devices recently provided for our new Vice Chancellor. Just about every IT service in the University will have a dependency somewhere on systems and infrastructure supported by Infrastructure Services. Even as you walk around Oxford city and travel beyond the ring road you will likely be walking over some of the 300km of fibre optic cabling that we maintain beneath the streets and highways.
We spoke to Dr Michael Fraser, Director of Infrastructure Services, to find out more.
Have you always been an IT infrastructure person?
No. Not at all. I arrived in Oxford in 1995 still completing a PhD that investigated the role of the fourth century Emperor Constantine in the building of what is now known as the Church of the Holy Sepulchre site in Jerusalem. My original ambition was to develop an academic career. However, lack of posts in Late Antiquity, and a lack of money, initially led me down a humanities computing route.
My first job in Oxford was a fixed-term contract post supporting humanities academics nationally to make more effective use of IT in teaching and research. Through a combination of opportunities and, it turned out, a skill in writing grant applications, I led or co-led a number of national projects and services which evolved from humanities to digital libraries, and then onto more fundamental IT infrastructure projects relating to open-source software, access management, online portals, and research e-infrastructure. It was perhaps inevitable that my next job would be managing, and then directing, IT infrastructure services (I just didn’t know it back then).
What is IT infrastructure?
It is everything you probably already think it is – network cables, data centres, switches and routers, servers and disks. It is also the platforms built on this kit – network management, Microsoft, Unix, virtualisation, and endpoint devices. The complexity lies in in how all these components are architected and interconnected into the systems on which these technical platforms and ultimately, the services everyone uses, are built. Today, of course, much of this infrastructure is virtualised as cloud computing, whether in a University data centre or one run by public cloud providers like Microsoft or Amazon. However remote and cloudy an IT service might seem, it will always depend on stacks of physical kit in data centres somewhere. Key to all of this is designing IT infrastructure that is reliable, and resilient to the inevitable equipment and software failures that arise.
IT infrastructure is not only in data centres. It is literally part of the fabric of buildings. The University has a large portfolio of new and refurbishment building projects and Infrastructure Services is often involved in the provision of network services to these buildings. We may also advise, with colleagues elsewhere in IT Services and Estates Services, on audio-visual deployments, access control and digital signage, and other IT components on which the daily operations of a building depend. The design and implementation of networks and audio-visual can be quite challenging for capital building projects. There are many inter-dependencies, and the overall IT budget may be set, as part of the overall build programme, years before the equipment is actually required. Interim changes to building design, as well as inflation, product availability, and procurement outside the main building programme, makes for additional complexity and risk. Fortunately, networks (and AV) are now rarely the last items on the schedule. Infrastructure Services is now involved in most significant building projects from the earliest stages in design and planning.
How is Infrastructure Services organised?
We are the largest grouping of staff in IT Services with over 100 people, the majority of whom have roles with significant technical expertise. Infrastructure Services is organised along generally logical lines, with groupings around networks and data centre services; technical platforms; research computing and support; and enterprise applications. Each group has a head supported by a number of teams. This structure allows for team-based working and encourages both intra- and inter-group collaboration. However, as services and technologies evolve we keep the group structure under review.
What are today’s challenges that you are addressing?
Responding to new ways of operating
IT infrastructure is not getting any simpler. Increasingly, a bit like cars, more of what used to be tangible metal and spinning disks, is now software-defined. Where once we had multiple servers and network switches dedicated to specific applications in a University data centre, now we have virtual servers and storage stretched across multiple data centres (which may not be University owned). Virtualised or cloud infrastructure is making it easier to increase capacity for a service, or to help ensure a service continues to run should a data centre disappear, but it has changed the knowledge and expertise required to use this infrastructure effectively and securely. It can be a challenge to maintain the services that the University cares about whilst developing skills in response to innovations and new ways of operating.
Managing data centres efficiently and effectively
Less than five years ago IT Services hosted infrastructure across four different data centres in Oxford, two of which were ‘legacy’, to put it kindly. Today, we host infrastructure across three data centres (one of which is ‘off site’ and one of which supports research computing) and are carefully considering how we can increase our use of public cloud providers. This strategic change in how we manage our services is reflected in the projects we have underway and in the changing roles within Infrastructure Services. We have not solved the data centre challenge though. One of the Digital Transformation projects to which we are contributing is looking at the next phase of research computing for the University and what (or who) may be required for hosting the next range of liquid-cooled high performance compute clusters. As you might imagine, a large chunk of the IT Services energy bill arises from our data centres so running the data centres and the kit therein as efficiently as possible is an ongoing concern.
Protecting University data
Helping to keep University data safe is one of our highest priorities. One of the key services we provide is data backup – from enterprise databases through to a postgraduate’s laptop. Increasing complexity can also result in greater risks around cybersecurity and the protection of data more generally. We work closely with the University’s information security and compliance teams in developing strategy and policy, and in helping to ensure our systems are designed and operated securely. There is almost always at least one IT project underway that is responding to evolving security threats.
Business continuity in emergencies
The energy crisis this winter, especially in relation to the risk of scheduled power outages, focussed significant effort on planning for continued access to IT services in the event of multiple University buildings losing power. The challenge was to ensure that buildings with power continued to have access to IT services, and that those services remained accessible remotely. Tracing end-to-end IT network and system dependencies from, e.g., a college building, through multiple network communications rooms, out to data centres, and back, identified the need for alternative solutions to be put in place to keep our services available. Implementing crisis management plans (in advance) requires agility and imagination - how we responded to the emerging pandemic in 2020 (in rapidly increasing VPN and remote desktop capacity and devising a way to efficiently provision Teams sites) is another example.
Managing change and decay
In the Admission of the Vice-Chancellor to the University, both the Chancellor and the newly elected Vice-Chancellor, in their respective addresses, referred to the quotation in Giuseppe Tomasi di Lampedusa’s The Leopard, "For things to remain the same, everything must change" (or evolve). If the University is to remain at the forefront of research, education, and public engagement, then many things must change and evolve, including IT, even if the changes are not initially popular with everyone.
How we manage identities and the multiple relationships people have with the University, beyond issuing a University Card, is one such challenge to address. Ensuring a consistent and good experience of eduroam wherever you might be in the University, is another. Parts of the University remain, even in a time of Teams and hybrid working, passionate about telephones; and everyone cares (or really should care) about data management and security. And it can be a long process to close down a service even as the underlying infrastructure (and the application data) becomes increasingly vulnerable to outages and security incidents. This year we will see the closure of SharePoint 2013, OxFile and, hopefully, WebLearn. Both identity management and Wi-Fi are included amongst the many Digital Transformation projects to which Infrastructure Services is making a leading contribution; we are also working on a plan for the next stage of University telephony that increases convergence with other communication tools whilst retaining telephone handsets for those who need them.
How is your work making a difference at Oxford University?
What matters is that we ensure that the services are maintained, available and reliable; and change in step with how the University requirements evolve. Infrastructure Services, the group, depends on our highly capable people – engineers, technical and team leaders – to operate and develop our services. We have, for example, improved the Wi-Fi service where we can and have a Digital Transformation activity to enable a step-change in how Wi-Fi is delivered across the University (including integration with city-wide 5G when available and sensible to do so). We have improved the availability of fundamental IT infrastructure, such as data centres and networks, to reduce the number of outages, whilst increasing the consolidation, virtualisation, and automation of our server and storage estate. We have projects underway to improve mobility and hybrid working, research data management, and to address the identity management challenge.
Ultimately, Infrastructure Services is an enabler – we enable other parts of the University, as well as IT Services, to deliver IT to their constituencies; and through the direct provision of IT services, we enable people to do their work across the collegiate University. So long as we are an enabler, and not a blocker, to students and staff making a difference, then we are doing our job.