Monash eResearch Centre
eResearch is an uncommon professional term. Thus this page will describe what a Director of eResearch does and my three most substantive impacts on the organisation.
In 2010 I was recruited into the newly reconstituted Monash eResearch Centre (MeRC) to provide vision and administrative know-how. eResearch, now commonly referred to as digital research infrastructure, is about enabling researcher-led technology. The Centre was Monash University’s strategic initiative to increase research performance by establishing a world-leading environment underpinned by digital research infrastructure. A decade later, the University’s growth in research rankings was among the top three globally. The Centre’s institutional and national activities have grown by $1m per year, and the Centre is now globally renowned.
In practice, the Centre is the trusted custodian of the Institution’s computing fuel for research (processing power, enduring data & digital transformation capability). I have pioneered an operating model that achieves flat-lined expenditure and very high user satisfaction, despite a sustained 75-85% CAGR demand in consumption growth. It has over 2000 users / Chief Investigators, and these users provide for an estimated 100,000 other researchers. It is ISO9001 certified, centred primarily on operating DevOps shared services, alignment with the institutional ecosystem and delivering co-designed projects.
As Deputy Director, my primary roles are driving vision, direct responsibility for half the initiatives (the fabric for digital infrastructure, cloud and the research data lifecycle strategies), coaching our next generation of leaders, implementing digital and data governance, and the Centre’s financial accountability.
Innovation requires an innovative culture and engaging with innovative partners. The following partnerships have transformed the University...
Established a globally renowned engineering centre
I pioneered the transition of cloud & HPC into the data processing era, developing and re-training staff on the journey. I positioned team members for leadership roles in the national and international communities. We developed a global brand; for example, “Nectar / Monash was the first to push HPC and GPUs on OpenStack. Whereas Monash started with OpenStack and then did HPC, Cambridge has done it vice-versa” – Cambridge University, 2019
I positioned the Centre as an early adopter of Mellanox’s ConnectX NICs, based on a technology (RDMA) that dominated the HPC market but was largely unknown to general computing. We produced the global first reference implementation of ConnectX for societally relevant use cases, launching at OpenStack Summit Tokyo 2015. By 2019 ConnectX is 70% of market share (for NICs above 10GbE).
Being one of Australia’s top universities, my obligatory “biggest” story is building the southern hemisphere’s largest Ceph storage cluster, 2016. However, my “proudest” account is when the ARDC partnership re-invented how clouds, security, machine learning, data science and AI make solving grand challenges easier, 2021.
Scaling out cybersecurity and data governance to 1000s of applications
I partnered with our CISO and the Australian global start-up Bugcrowd on a crowdsourced approach to identifying vulnerabilities in our ecosystem of hundreds of advanced research-led digital systems. I introduced scalable business processes to incentivise researchers to participate and respond to these vulnerabilities proactively. Case Study: Monash University Leverages Bugcrowd for improved security visibility and continuous assurance, 2021.
I subsequently partnered with Nvidia and the ARDC to explore the role of DPUs and AI in cybersecurity. Our approach has always been frictionless and zero-trust. However, the technologies and processes need to evolve as the risk tolerance tightens. Else the friction will increase. In this partnership, we use the research sector as a microcosm of a connected society to explore the transition to security everywhere.
The amount of cybersecurity work to satisfy emerging best practices for the typical IT use case will limit a researcher’s liberty to collaborate. Distributing that work adds the substantive and scalable resource for more emergent and democratised ways to connect. It lowers the need for researchers to change their software due to increasing cybersecurity needs and decreases the percentage of precious computational resources (CPUs and GPUs) used for security. However, these technologies are not yet mainstream (let alone GA’d), and part of my role is to drive us all there!
Changing IT culture - pioneering DevOps beyond cloud teams
In collaboration with Red Hat, I pioneered the adoption of Ceph (software-defined storage) to contend with the breadth in heterogeneity and scale of storage needs. It drove a transition in our capabilities towards a DevOps methodology for deployment and continuous improvement for storage. We ultimately won RedHat’s storage client of the year award in 2017.
In partnership with the University’s networking team and Cumulus Networks (acquired by Mellanox and then Nvidia), I pioneered DevOps adoption by our and the broader Institution’s networking capability. Accolades include the early adoption of Cumulus and accelerating Web-Scale adoption.
However, from a researcher’s lens, pioneering DevOps adoption in high-performance computing facilities has made the most significant impact. We bought the cloud and HPC worlds together at Supercomputing 2016 and my team made substantial contributions to the adoption of OpenStack for scientific research, 2016