The document summarizes the UK's investments in e-infrastructure for research from 2011-2015. It discusses the major investments made in high performance computing (HPC), networking infrastructure, and big data projects. The investments totaled £160 million in 2011-2012, £189 million in 2012-2013, and £257 million in 2014-2015. It also summarizes the results of a survey of the UK's e-infrastructure, including details on the largest HPC systems and datasets. Finally, it mentions that the Research Councils UK (RCUK) developed a roadmap for the UK's e-infrastructure with a vision for an integrated infrastructure to support researchers.
Asymmetry in the atmosphere of the ultra-hot Jupiter WASP-76 b
UK e-Infrastructure for Research - UK/USA HPC Workshop, Oxford, July 2015
1. UK e-Infrastructure for Research
Michael Ball, BBSRC
Frances Collingborn, NERC
Martin Hamilton, Jisc
David de Roure, ESRC / University of Oxford
Photo credit: STFC 1UKUSAHPC - July 201509/07/2015
2. UK e-Infrastructure for research
1. UK e-Infrastructure for research
– Public funding for major science facilities and institutes
– Support for translation from R&D into business
2. e-Infrastructure survey
– Build inventory of the e-Infrastructure
– Operating systems and software environment
– Funding and budgeting models
– Training and support arrangements
– Academic and industrial impact
3. RCUK e-Infrastructure roadmap
– Vision and aspirations
– Investment plan
Photo credit EPCC / EPSRC
4. UK e-Infrastructure for research
HPC Project RC Amount/£M
National Service EPSRC, NERC 43
Hartree Centre STFC 30
DIRAC STFC 15
GRIDPP STFC 3
The Genome Analysis Centre
(TGAC)
BBSRC 8
Monsoon NERC/Met Office 1
JASMIN2 & CEMS NERC, & UKSA 7.75
Regional Centres: N8, SES5,
MID+, HPC Midlands, ARCHIE-
WeSt
EPSRC 6.5
JANET Network and
Authentication Moonshot
Jisc 31
HPC Data Storage EPSRC, STFC 15
Total 160
Investments by BIS, the Research Councils and HEIs
have resulted in core elements of the national e-
Infrastructure being put in place.
» 2011-2012 - £160m
Investments were made in core HPC and
Networking infrastructure. In addition investments
were made in the Authentication Infrastructure
Moonshot (now known as Jisc Assent).
» 2012-2013 - £189m
» 2014-2015 - £257m
6. UK e-Infrastructure for research
Big Data Project RC Amount/£M
Digital transformations in arts and
humanities
AHRC 8
E-infrastructure for biosciences BBSRC 13
Research data facility and software
Development
EPSRC 8
Administrative data centres ESRC 36
Understanding populations ESRC 12
Business datasafe ESRC 14
Biomedical informatics MRC 55
Environmental virtual observatory NERC 13
Square Kilometre Array STFC 11
Energy Efficiency Computing
Hartree Centre
STFC 19
Total 189
Investments by BIS, the Research Councils and HEIs
have resulted in core elements of the national e-
Infrastructure being put in place.
» 2011-2012 - £160m
» 2012-2013 - £189m
Big Data projects using funds announced by the
Government in December 2012 were funded at this
time. Major Awards have been made to 18 centres
in the UK, 16 of whom are HEIs.The pre-eminent
role of HEIs in managing and providing national
and Large Specialist data and compute services to
UK academia is emphasised by these awards.
» 2014-2015 - £257m
7. UK e-Infrastructure for research
Energy Efficient
Computing
Infrastructure
(STFC)
De-identified admin
(including health) data
Busines
s data
Open data
(public
sector)
Social media
data
Researc
h data
Longitudin
al survey
data
Open data
Securely held data
Environme
nt data
Business Datasafe
(ESRC)
Admin Data Research
Centres (ESRC)
High Performance
Data Environment
(NERC)
Clinical
data
Medical Bioinformatics (MRC)
Understanding Populations
(ESRC)
Clinical Practice Datalink
(MHRA, NIHR)
100,000 Genome Project NHS)
Research Data Facility (EPSRC)
European Bioinformatics
Institute (EMBL)
Bioscience E-Infrastructure
(BBSRC)
Square Kilometre Array (STFC)
Digital Transformations
(AHRC)
Archive
data
Open Data
Institute
Commercial
Research
Understanding
Populations (ESRC)
RCUK Big Data
21st century raw material
9. UK e-Infrastructure for research
Investments by BIS, the Research Councils and HEIs
have resulted in core elements of the national e-
Infrastructure being put in place.
» 2011-2012 - £160m
» 2012-2013 - £189m
» 2014-2015 - £257m
Three major investments dominated this period:
» Centre for Cognitive Computing at the Hartree Centre.
This was funded at the £115M level with a further £230M from IBM
» A 10 Pflop Supercomputer for the Met Office (£100M)
» AlanTuring Centre for Data Science (£42M)
In addition it was announced that a further £100M would be made available to the SKA Project as
part of Big Data Investments.
Photo credit: EPSRC
10. April 2015 BBSRC bioscience big data infrastructure funding:
» £1.79M to build a next generation image repository, to make available original scientific image
data that underpins life sciences research.
» £2M for big data infrastructure for crop genomics, stimulating new opportunities in crop
development to help improve some of the world's most important crops.
» £1.9M to establish infrastructure for functional annotation of farmed animal genomes, to
help feed us in the future by providing an important framework for the discovery of genetic
variation in domesticated animals and how that influences their characteristics.
» £1.78M to create cyber infrastructure for the plant sciences.The UK iPlant node that will help
to spread expertise and best practice between the UK and US. UK/US collaboration with
University of Arizona and theTexas Advanced Computing Center.
UK e-Infrastructure for research
11. UK e-Infrastructure for research
bit.ly/dowlingreport bit.ly/bis8great
Context:
› Reviews, e.g. Pearce,
Diamond, Dowling,
Shadbolt
– Demonstrable efficiency,
effectiveness and
productivity
› UK Government
Industrial Strategy
– 8 GreatTechnologies
– CatapultCentres
› Cultural shifts
– Open Science
– OpenAccess
– Open Research Data
12. UK e-Infrastructure for research
bit.ly/hauserreportbit.ly/jischpc
Drivers:
› Shared facilities and
industry access
– Finding them
– Using them (kit & people)
› Big push for translation
and consolidation
– New Catapult Centres
– Farr Institute,
FrancisCrick Institute,
AlanTuring Institute
› Impact of Austerity 2.0
– Comprehensive Spending
Review, Autumn 2015
14. e-Infrastructure Survey
What we did:
› Build an inventory of UK research e-Infrastructure
– Including interconnects, storage, accelerators etc
– Gathering data on use of cloud technologies
› Itemize operating environment
– e.g. OS distributions, schedulers, filesystems, authentication
& authorization
› Funding and budgeting models
– Power costs, PUE, split between CAPEX/OPEX,
location of scientific computing in the institution
› Training and support arrangements
– Where is support effort spent, role of women in HPC
› Academic and industrial impact
– Grants, papers, businesses using the facilities
Photo credit: CC-BY HPC Midlands
20. e-Infrastructure Survey
› Top 9
Large &
Specialist
(by number
of users)
1. Large and Specialist Services
Organisation name System name
What are the top three research
areas the system is used for?
Total number
of processor
cores in the
system
Total usable
storage for
HPC users
(TB)
Number of
registered
users
Theoretical
Peak
Performance
(Tflop/s)
NERC (operated by STFC) JASMIN
Climate Science, Earth
Observation, environmental
genomics 4,500 25 Over 10,000
STFC Hartree Centre Blue Wonder
Modelling & Simulation (CFD,
Materials, and Computer Aided
Formulation) 24,000 9000 750 - 1,000 200
Norwich Bioscience Institutes
(TGAC, JIC, IFR, TSL)
Bioinformatics, mathematical
modelling. 9,000 4,000 750 - 1,000
DiRAC @ University of
Cambridge (HPCS) Darwin
Life Sciences. Atomic structure.
Computational Fluid Dynamics. 9,600 2,847 750 - 1,000 200
STFC Scientific Computing
Division
UK e-Science Certification
Authority
Supports all UK research. Major
users Particle Physics 750 - 1,000
STFC Scientific Computing
Division SCARF
Computational Chemistry Plasma
Physics, Processing Satellite
images Support of ISIS, CLF,
RAPSP, DLS user communities 7,000 320 500 - 750 165
STFC Hartree Centre Blue Joule
Modelling & Simulation (CFD,
Materials, and Computer Aided
Formulation) 98,000 6000 200 - 500 1,200
EMBL-EBI - European
Bioinformatics Institute Embassy Cloud Life science research 31,000 3,200 200 - 500
DiRAC @ EPCC DIRAC BG/Q QCD, Soft Matter Physics 98,304 1,000 200 - 500 1,258
21. e-Infrastructure Survey
› Top 9
Large &
Specialist
(by number
of users)
1. Large and Specialist Services
Organisation name System name
What are the top three research
areas the system is used for?
Total number
of processor
cores in the
system
Total usable
storage for
HPC users
(TB)
Number of
registered
users
Theoretical
Peak
Performance
(Tflop/s)
NERC (operated by STFC) JASMIN
Climate Science, Earth
Observation, environmental
genomics 4,500 25 Over 10,000
STFC Hartree Centre Blue Wonder
Modelling & Simulation (CFD,
Materials, and Computer Aided
Formulation) 24,000 9000 750 - 1,000 200
Norwich Bioscience Institutes
(TGAC, JIC, IFR, TSL)
Bioinformatics, mathematical
modelling. 9,000 4,000 750 - 1,000
DiRAC @ University of
Cambridge (HPCS) Darwin
Life Sciences. Atomic structure.
Computational Fluid Dynamics. 9,600 2,847 750 - 1,000 200
STFC Scientific Computing
Division
UK e-Science Certification
Authority
Supports all UK research. Major
users Particle Physics 750 - 1,000
STFC Scientific Computing
Division SCARF
Computational Chemistry Plasma
Physics, Processing Satellite
images Support of ISIS, CLF,
RAPSP, DLS user communities 7,000 320 500 - 750 165
STFC Hartree Centre Blue Joule
Modelling & Simulation (CFD,
Materials, and Computer Aided
Formulation) 98,000 6000 200 - 500 1,200
EMBL-EBI - European
Bioinformatics Institute Embassy Cloud Life science research 31,000 3,200 200 - 500
DiRAC @ EPCC DIRAC BG/Q QCD, Soft Matter Physics 98,304 1,000 200 - 500 1,258
22. e-Infrastructure Survey
› Regional centres (by total cores)
2. Regional Systems
Organisation name System name
What are the top three research areas
the system is used for?
Total number of
processor cores
in the system
Total usable
storage for HPC
users (TB)
Number of
registered users
Theoretical
Peak
Performance
(Tflop/s)
HPC Wales Various (distributed system)
Advanced Materials & Manufacturing,
Life Sciences and Energy &
Environment 16,816 702 2,000 - 5,000 319
N8HPC Polaris 5,312 175 200 - 500 138
ARCHIE-WeSt ARCHIE
Molecular dynamics, CFD, Plasma
Physics 3,920 148 200 - 500 38
HPC Midlands Hera
Advanced Materials Energy Efficient
Transport 3,008 120 100 - 200 48
23. e-Infrastructure Survey
› Regional centres (by total cores)
2. Regional Systems
Organisation name System name
What are the top three research areas
the system is used for?
Total number of
processor cores
in the system
Total usable
storage for HPC
users (TB)
Number of
registered users
Theoretical
Peak
Performance
(Tflop/s)
HPC Wales Various (distributed system)
Advanced Materials & Manufacturing,
Life Sciences and Energy &
Environment 16,816 702 2,000 - 5,000 319
N8HPC Polaris 5,312 175 200 - 500 138
ARCHIE-WeSt ARCHIE
Molecular dynamics, CFD, Plasma
Physics 3,920 148 200 - 500 38
HPC Midlands Hera
Advanced Materials Energy Efficient
Transport 3,008 120 100 - 200 48
24. e-Infrastructure Survey
› Top 8 HEIs
(by total
cores)
3. HEI Systems
Organisation name System name
What are the top three
research areas the system is
used for?
Total
number of
processor
cores in
the
system
Total
usable
storage
for HPC
users (TB)
Number of
registered
users
Theoretica
l Peak
Performan
ce
(Tflop/s)
Imperial College London cx1 21,558 2,000 750 - 1,000
University of Bristol BlueCrystal
Chemistry, Aerospace Eng,
Geographical Sciences 9,000 740 750 - 1,000 240
University College
London Legion
Chemistry, Physics, Biological
Sciences (according to REF
Categories) 7,816 356 500 - 750 115
Imperial College London cx2 7,000 500 0 - 100 60
University of
Manchester
Computational Shared
Facility
Computational Chemistry /
MD CFD FEA 6,288 750 750 - 1,000 111
Durham University Hamilton
Condensed Matter Molecular
Dynamics Fluid Dynamics 5,600 350 200 - 500 75
University of Oxford Arcus-B 5,440 432
2,000 -
5,000 538
Lancaster University HEC (High End Cluster)
High Energy Physics
Condensed Matter Theory
CFD 4,784 1,530 200 - 500
31. RCUK e-Infrastructure roadmap
From the roadmap document:
“Our aspiration is for the UK to have an integrated e-
infrastructure: one that is run and managed as a whole without
silos or boundaries, where there are simple processes by which
users can get access to the e-infrastructure they need across the
eco-system, as appropriate for the type or stage of research
they are doing.We need to consider how best to integrate:
» Vertically up and down the eco-system pyramid, so users
have easy access to the most appropriate type of e-
infrastructure they need;
» Horizontally across the different elements, as shown in the
diagram;
» Across the different research communities and the
different stakeholders;
» Internationally, across other national e-infrastructures to
deliver end-to-end services in the global environment of
collaborative research.”
bit.ly/eroadmap
33. That’s all, folks…
33
Except where otherwise noted, this
work is licensed under CC-BY
Martin Hamilton
Futurist, Jisc, London
@martin_hamilton
martin.hamilton@jisc.ac.uk
UKUSAHPC - July 201509/07/2015