Researchers based in Western Australia can no longer directly access nationally funded resources on the Nectar Research Cloud. In an email sent to researchers in WA, Nectar announced that this decision was taken because the Pawsey Supercomputing Centre, originally providing support to Nectar in WA, had “decided to discontinue providing resources to the Nectar Federation”.
Although the Nectar Research Cloud is nationally funded as part of the National Collaborative Research Infrastructure Strategy (NCRIS), Nectar has decided to close the door to WA researchers because of “the implications of the absence of a WA-based resource provider and co-investment from WA research institutions”.
This move has come as a surprise to institutions in WA, who as far as I have been able to tell, did not know that they were about to be cut off.
The decision raises critical questions about how an entire state of Australia can be cut off from nationally funded research infrastructure?
What is the Nectar Research Cloud?
Nectar (National eResearch Collaboration Tools and Resources) was established in 2009 with funding from the National Collaborative Research Infrastructure Strategy (NCRIS), which has invested A$2.8 billion since 2004 in research infrastructure. The original funding for Nectar was A$47 million from the Super Science Initiative and A$9.4 million under the NCRIS strategy 2013 program.
In total, Nectar has received A$61 million in government funding matched by co-investment of A$54 million from Australian universities and research organisations.
The remit of the organisation was to create eResearch tools, virtual laboratories, a research cloud and a secure and robust hosting service. The research cloud part is essentially a service similar to the basic functionality of cloud providers like Amazon Web Services, Google Cloud, Microsoft Azure and others.
Researchers could apply for access to computing resources in the cloud at no direct cost to them on a merit basis.
Some uses of Nectar resources have resulted in what are called “Virtual Laboratories”. These virtual laboratories host data and tools from disciplines such as genomics, astronomy, marine science and climate research.
Why not use commercial resources?
The big question regarding governments funding these types of infrastructure projects is why they would spend so much money when a range of commercial companies do the job already and one could argue, do the job better?
One immediate issue with funding computer infrastructure is that it only has a limited lifetime of 3 to 4 years before it needs completely replacing. Whilst this may be acceptable for extremely specialised super computers that are essential to certain types of research, the advantages for regular server computing resources is not so clear.
The field of cloud computing is evolving at an extremely rapid pace. Amazon is adding services to its cloud offerings on a monthly basis. It provides around 120 different cloud services that can be paid for on an hourly, monthly or yearly basis. In contrast, the Nectar equivalent functionality as part of its computer cloud operation only offers the basic services of a virtual machine and storage.
Whereas commercial cloud providers are investing billions of dollars into new services, especially in the machine learning area, government funded infrastructure of this sort is simply not able to compete.
It is not clear how an organisation that has been funded to provide a national service can unilaterally decide not to provide Western Australia with those resources. Nectar is in the portfolio of the Federal Government’s Department of Education and Training and so ultimately, it would be their responsibility to ensure that resources funded nationally are available to all researchers in Australia.
The suggestion by Nectar for researchers in WA was to ask another researcher in another state to apply for resources, negotiate or pay for access to a facility in another state, or “consider using a commercial cloud provider” and again pay for that service.
If that is indeed the case, it would be good if WA could get back its share of the A$61 million to spend on these options.
In preparing this post, I reached out to both the Pawsey Supercomputing Centre and Nectar for comment but had not received a reply by the time of posting. I will update if and when I receive a reply.
Update: 5th September 2017 15:33 AWST
Representatives from both Pawsey Supercomputing Centre and Nectar have now replied to this article and are preparing statements of their perspective on events and the current situation with regard to WA researchers and the Nectar cloud.
Authors: David Glance, Director of UWA Centre for Software Practice, University of Western Australia