The below text was used as information to prompt ideas ahead of our sandpit events in May 2022.
A series of workshops will be organised within the scoping project in order to…
- Broaden discussion and gather perspectives from a wide range of sources
- Capture evidence to support the roadmap
- Deliver recommendations
The hosting of workshops will be outsourced in order to engage a range of organisations in the process. A budget of £210k is reserved to fund this activity.
Questions to address
All workshops should address these main questions:
- What is the scale of carbon reductions possible by 2030, 2035 and 2040?
- Where will these carbon savings come from? Are there net financial costs or benefits? Are there additional costs or benefits in, for example, use of users’ time?
- What decisions need to be taken by 2025, 2030, 2035?
Evidence could take the form of measured differences of energy use in different scenarios, literature review, expert opinion or user feedback. Projects should justify the type of evidence used and its scope, and show how the evidence adds to existing published sources of evidence. Evidence from surveys may be combined with evidence based on practical examples.
Workshops were held online in order to enable broad participation, including international participation where possible, without incurring the expense and carbon footprints of travel. Creative approaches to promoting active discussion and engagement online will be particularly welcome.
The programme as a whole will give a broad community view on how carbon-lifecycle costs of data infrastructure can be managed, covering issues from the choice of building materials for a machine room down to the factors affecting the decisions that users make and the efficiency of their interactions with the infrastructure.
Assessment of Proposals - The Three Es
Proposals will be judged on three criteria:
- Excellence: does the proposed approach have a clear plan for creating and running events which will convey the latest published research and operational thinking on each topic?
- Evidence: will the proposed event(s) produce authoritative evidence and recommendations that can be used to underpin recommendations for the UKRI Net Zero DRI roadmap?
- Effectiveness: is there a clear plan for the event(s)? Have the proposal team thought through the process well enough? Is there creative use of online meeting and discussion options?
Potential Concepts for Discussion
The following list of potential topics is not intended to be exclusive, but to give an idea of the scope of ideas that will be considered. Topics may be combined if proposers argue that they are better handled together.
- Machine room and hardware: construction, cooling and power supply
- Review of best practice in efficiency of building construction (embodied carbon) and operations (e.g. Power Usage Effectiveness - PUE). Considerations of energy supply, including use of local renewables coupled with storage.
- Improving resilience : as DRI becomes central to more and more activities there is an increased need for continuity of service.
- Machine architecture and operation
- Carbon footprint of hardware (Product Carbon Footprint - PCF).
- Green computing, e.g. system-on-chip designs (e.g. ARM);
- Are there options for improving the handling of end-of-life ICT equipment?
- Consolidation of services to enhance efficiency in use of carbon and other resources
- Can cloud services and other multi-use data centres both enhance the carbon efficiency of digital infrastructure and improve the quality of service for users? How are services and technologies going to evolve between now and 2040? Given that idle infrastructure represents wasted carbon invested in the footprint of buildings and machines, what can cloud contribute to optimising the utilisation factor (UF)?
- Effective Computation: making each simulation count
- When the racks in the machine room are running at full power running scientific software, are they doing useful work? What if John has a bright idea at 5pm on Friday and quickly sends off a set of jobs, and then finds on Monday the 8000 core hours need to be repeated because he mis-typed one number? What if Jane has her hard drive full of data that she wants to analyse, but never finds time? How can we help them to reduce the carbon footprint of their work? What is the potential for reducing the resource evaporation factor (REF), the amount of resources which get used but do not contribute to any research outputs? Is there a role for Open Science, as applied to scientific software development, in increasing efficiency?
- Feedback to users about their energy usage (e.g. this month, you used xxxMW.h for yyy jobs)
- And to PIs of projects, about their project as a whole
- And translate this to meaningful things (eg that is enough energy to run an average UK 2-person household for xxx months)
- Efficient management of data products
- Once a significant piece of work has been completed, will other scientists be able to re-use the results, or will they find it easier to re-run the simulation? What is the role of FAIR data principles in reducing redundancy (REDU)?
- Understanding the options for carbon capture from now to 2040
- What are the relative merits and risks of capturing carbon in the biosphere (e.g. soil and trees) versus geological capture? What are acceptable standards for durable carbon storage, what risk of reversal can be tolerated, and how should that risk be calculated? Does the REDD process (Reducing Emissions from Deforestation and forest Degradation) offer a model for establishing standards? Is there a need for research in durable storage systems? Can UKRI contribute to establishing global standards?
- The plus side of the digital research infrastructure carbon account
- How can DRI contribute to the broader societal transition to Net Zero targets? Consider the role of digitization of information, reduction of movement and transport, making industries and lifestyles more efficient. What is the role of Digital Twins?
- Net Zero: What holds us back?
- Looking across the whole range of activities, from major infrastructure investments down to personal choices by research students, what are the barriers to adoption of carbon efficient choices? Consider investment purchasing choices of major national infrastructure, institutional, departmental and also purchase by smaller teams and individuals. This workshop will take a holistic view, looking at financial rules, ways of working, and potential disruptions that discourage change.
- Encouragement through recognition is currently unclear, contributing to the transition to sustainability is not a specific research evaluation metric.
- Overall approach: structure of the roadmap and the assessment of carbon impact of the DRI
- [this topic could be handled as a series of meetings run by the core group, rather than as a single workshop]
- Metrics and Indicators of carbon footprint and efficiency
- Handling of decision points
- Integration with wider aspects of sustainability policy at UKRI and institutions hosting facilities (e.g. policies on electricity supply and sustainable procurement)
- .. other cross-cutting issues