UCT eResearch is about IT-enabled research discovery at the University of Cape Town. This slide show was presented to IT faculty managers and technical staff to demonstrate how UCT eResearch can assist faculties at the university to raise eResearch capacity.
1. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
eResearch UCT
Anthony Beitz
Director UCT eResearch
www.eresearch.uct.ac.za
2. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Agenda
1. Introduction to eResearch
2. eResearch @UCT
3. Examples of eResearch at UCT
4. Accessible capabilities & infrastructure
3. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Emergence of Research Methodologies
4. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
eResearch is 21st century research
discovery through the application
of advanced computing and IT
5. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
It is transformational research:
Second image courtesy of Dr. Govinda Poudel, with permission.
Third image courtesy of Electronic Visualization Laboratory & Dept. of
Psychiatry, UIC, with permission. Photo by Philip Chan, Monash Uni.
6. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
It is accelerated research:
Brains of ~100 study subjects (controls,
pre-symptomatic and Huntington’s
Disease sufferers) visualised at once!
Data: N. Georgiou-Karistianis (Monash Uni.) Visualisation: Y.
Benovitski (Monash Uni.)
7. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
MRI Scanner
Collect
GPU Cluster
Compute
CAVE2TM
Comprehend
eResearch version of a microscope
8. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Research Data Commons
Communicate
Virtual Lab
Collaborate
eResearch version of a microscope
Compute
Collect
Comprehend
9. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Research Data Deluge
Data generation rates:
• Next Generation Sequencer > 200 TB / year
• Direct electron detectors > 350 TB / year
(equivalent to synchrotron beamlines)
• Large Hadron Collider 15 PB / year
• Square Kilometer Array 0.3-1.5 EB / year
Laboratory
based
Large
facilities
10. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Research Acceleration & Transformation
IMPLICATIONS
Need to grow
access to
infrastructure,
capability and
tools
Pace and scale
increasing
• Technology is accelerating the
pace and scale of research
• Large-scale data requires more
structured approach data
management approach
KEY CHANGES
Expert guidance
and training to
leverage new
approaches
New tools and
methodologies
• New and more powerful
instruments
• Digital recognition of text,
speech and imagery
• Crowdsourcing and citizen
science
Swift capability to
establish and
support
collaborations
Internationalisation
• Research funding is increasingly
based on strong international
collaborations
• Internationalisation is
introducing new complexities in
health, ecology and economy
(epidemics, climate change,
Global Fin. Crisis)Slide courtesy of Nathan Bailey & Lisa March (iSimplify)
11. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
“Scholarly practices across an astounding wide range of disciplines have become
profoundly and irrevocably changed by the application of advanced information
technology. This collection of new and emergent scholarly practices was first
widely recognized in the science and engineering disciplines….humanities
continue to showcase some of the most creative and transformative examples
of the use of information technology to create new scholarship.”
12. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
http://www.canberra.edu.au/research/ucresearch/e-research
13. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Transforming research practice
• Unsuccessful structure solution
of PlyC structure
• However, raw data archived in
MyTardis, the Monash bioscience
research data repository
• New technological
advancements in 2010 and easy
access to the archival data led to
the PlyC structure being solved
Monash’s Dr. Sheena McGowan (under the direction of Prof.
James Whisstock and Assoc. Prof. Ashley Buckle)
PlyC
http://www.quantumdiaries.org/
14. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
About Researchers
• Research is about “doing things we don’t know how to do”
They work in an interpretive mode
– Characteristics include: an iterative process; open-ended; and
thrive on ambiguity
– Service providers try to remove ambiguity = breakdown in
communication and understanding
• Their requirements and goals will change over time, and very
likely won’t be well known at the start
• Generally more loyal to their research community than their
institution
• May require an ICT capability for only a short period of time
• Very resourceful and driven by research outcomes
15. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Creative Solutions
• Solutions are not always
“designed/engineered” – they are often
“grown” via an iterative process
• If researchers engage all the way along the
iterative process -- they will very likely engage
at the end
• The research challenges are unique and
therefore the solutions are often “unique” -
no one size fits all
Slide courtesy of Prof. Paul Bonnington (Monash Uni)
16. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Engaging with Researchers
• It’s basically about the relationship and demonstrating that
you are listening and delivering ongoing value
• Never push solutions or technologies
• Partner with Researchers and go on the journey of the
unknown together
• Solutions should be led and owned by the Researcher
– Agile methodologies and mindsets lend themselves to this
outcome
• Work in with their community
• Reduce the barriers
• The three T’s of an innovation environment: Talent,
Technology and Tolerance (for ambiguity and risk)
17. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
18. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
History of UCT eResearch
Began with UCT’s ICT Services’ (ICTS) Executive Director, Sakkie
Janse van Rensburg, searching for a better way to support
researchers
2009: UCT ICTS staff trained in HPC
2010: Built 1st cluster and visited Australia to investigate
eResearch
2011: HPC-trained staff focused on researchers, UCT researcher
appointed to work with ICTS
2013: Best-of-breed eResearch model identified, global search
2014/Q1: Core partnerships and eResearch working group
formed
2014/Q2: Business case developed, buy-in by UCT’s Senior
Leadership Group attained, initial eResearch team formed
2014/Q3: Strategic plan developed
2014/Q4: Interim director appointed
2015/Q1: Governance established
19. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
eResearch Responsibilities
Legend
A – Accountable
C - Consulted
I – Informed
20. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
eResearch
Centre
Research Office
•Advocacy
•Collaboration
ICTS
•Data Storage
•Compute
•Cloud
•Networks
DVC-Research
•Leadership
•Governance
•Strategy
•Resourcing
The Library
•Research visibility &
discovery
•Research data
management
Faculties
•Various capabilities
Researchers
21. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
UCT eResearch partners with research groups to accelerate and
transform research, connecting them to the most appropriate
hardware, software and services to sustain that capability
Collect Compute Comprehend
Collaborate Communicate Customise
Creating a world-class environment for
21st century discovery
22. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
As a partner and broker UCT eResearch will connect researchers with
world-class expertise and facilities (local or external) in:
• Collection and management of research data
• Modelling, simulation and data processing through HPC
• Comprehension of big data through visualisation and data science
techniques
• Promotion of collaborative research through virtual labs and cloud
resources
• Dissemination of research outcomes (including data and workflows)
• Development of customised research software, hardware and
services
23. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Resource Limits
• Requirements for advanced ICT
research infrastructure are huge
and growing
• Advice and fundamental
infrastructure are freely available
• For needs beyond fundamental
infrastructure, UCT researchers
engage early with the eResearch
Centre to estimate required funding
and to incorporate it into a grant
Imag: permabit.com/data-affordability-gap/
24. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
UCT eResearch does not
want to try to be the
ultimate Swiss army knife
for researchers
25. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Target Structure for UCT’s eResearch Centre
eResearch Centre
Steering
Committee
Directorate
Researcher Liaisons
Research
Infrastructure
Projects
Fledgling Research
Technology
Platforms
Virtual Labs &
Collaborative
Environments
Human Capital
Development
DVC-R
Exec. Dir. Research
Office
NB: The lines are not indicative of line reporting
Exec. Dir. ICTS
University
Librarian
26. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Functions
• Directorate
– Strategy, leadership, facilitate organisational change, comms and admin
• Researcher Liaisons
– Act as brokers between researchers and capability providers
• Research infrastructure projects
– Plan & execute research infrastructure projects
• Fledgling Research Technology Platforms
– Provide skilled resources, specialised infrastructure and relevant training
• Virtual research labs & collaborative environments
– Provide interim support for virtual research labs and collaborative envs
• Human Capital Development
– Co-ordinate staff and researcher training in new research cyberinfrastructure
27. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
28. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
29. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
SKA - answering some of science’s biggest questions
• One of the largest scientific
projects ever undertaken to
answer questions such as:
• What is dark energy?
• Was Einstein right about
gravity?
• Is there life elsewhere in
the universe?
• eResearch is supporting
storage, compute,
development of new
visualisation tool
• Exascale arriving soon…
www.skatelescope.org
30. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Enabling South African CERN researchers to
partner with their international colleagues
• CERN made up of a massive globally
dispersed community of researchers
• Successful international research
collaborations are dependent on a
good video conferencing experience
• South African researchers had a poor
experience
• UCT eResearch connected CERN with
national service providers to bring
video conferencing up to an good
level
31. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
• Key research projects e.g. SKA
and space geodesy
• Use of HPC facilities have tripled
in the past year
• 300 researchers across campus
assisted by HPC
• 33 citations (and counting)
Image: Stephen Williams Photography
High Performance Computing
32. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Facility Support: Olympus Virtual Microscope
The old: slide boxes
The new: Olympus VS120 Microscope
eResearch:
• Worked with the UCT’s
Clinical Lab Services
department to set up the
servers and the database
• Configured the system for
various access scenarios
including student,
departmental and public
access
33. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Facility Support: Centre for Imaging and Analysis
Provides the instrumental resources and expertise for structural analysis in all disciplines.
acquisition processing interpretation dissemination
Workflow:
34. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Facility Support:
• Nikon fluorescent microscope for
live cell imaging
• eResearch supported storage,
remote data analysis and
collaboration
Clinical Research CentreMolecular and Cell Biology
Tissue Culture Facility
• Clinical, laboratory and data
management systems needed to
pass an FDA or EMA audit.
• Software and computational
requirements met by eResearch
35. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Human Heredity and Health
in Africa
• First Pan African Bioinformatics
Network in Africa: H3ABioNet &
H3Africa Consortium
• UCT eResearch Centre, in collaboration
with Information and Communication
Technology Services (ICTS) and the
Computational Biology Group, will be
developing a data storage archive
solution
• The data will include genomic data and
access will be controlled according to
internationally accepted standards.
36. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
37. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Current capabilities & infrastructure
• Training
– HPC (beginner & advanced levels)
– Software carpentry (Python,R, Git, ..)
• Research data storage
• High Performance Computing
• Virtual machines and databases
• Project Managers, BAs, Devs, … for research projects
(through cost recovery)
• Assistance in research grant proposals (research
infrastructure)
38. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Ashley Rustin
ashley.rustin@uct.ac.za
Wednesday, 13 May
2015
39. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
UCT Core Network & Data Centres
40. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
UCT Bandwidth
• 1.2 Gbps International on WACS
• 10 Gbps Local via SANREN
41. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
UCT High Performance Computing (HPC)
1. Hex Cluster Nodes
• Dell C6145 x 9 = 18 nodes
• 4 sockets with 16 cores each
• 1152 cores
• 2.2TB RAM
2. HPC Cluster Nodes
• HPBL460c x 19 = 19 nodes
• 8 – 12 cores each
• 176 cores
• 648 GB RAM
Mellanox Infiniband 6036 SX
Interconnect
Dell C6145 compute Node
HP BL460C compute Nodes
42. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
UCT High Performance Computing (HPC)
3. GPU Nodes
• SuperMicro M2090 x 3 = 3 nodes
• 4 sockets with 16 cores each
• Nvidia TeslaM2090 & K40M (9856 GPU
cores)
• 36 CPU cores
• 192 GB RAM
4. High Memory Nodes
• Dell R820 x 2 = 2 nodes
• 8 – 12 cores each
• 176 cores
• 648 GB RAM
Dell C6145 & R820 compute Nodes
43. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
UCT HPC Storage
1. /home
• 2 TB Storage (NFS)
2. /scratch (old)
• 21 TB Storage (NFS)
3. /scratch (high speed)
• 92 TB (BGFS)
BackBlaze Storage Pod
Dell MD3060 Storage unit
44. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
UCT Research Data Storage
45. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
UCT Research Data Storage Benefits:
• Highly discounted for research – R3450 per TB or R34 500 per 10 TB
• Once-off payment
• Data is backup & replicated to UFS & UWC data centres
• Access your data from anywhere in the world
• Collect & store data from your instruments
directly to the central research data store.
• Contact: - eresearch@uct.ac.za
46. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Coming capabilities and infrastructure
• New visualisation tool & platform
• Federated access
• New data centre
• New compute cluster – 1000+ cores
• Training in data science, research data
management (in league with the Library)
• African Research Cloud (ARC)
47. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
48. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
Monash eResearch Centre
• eResearch is 21st century research discovery through the application of advanced
computing and IT
• UCT eResearch partners with research groups to accelerate and transform research,
connecting them to the most appropriate hardware, software and services to sustain
that capability
Collect Compute Comprehend
Collaborate Communicate Customise
49. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
50. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
The Team
Anthony Beitz (Director)
Timothy Carr
Jenny Dixon (Admin)
Piet Helm
Heine de Jager
Andrew Lewis
Chris Mtshengu
Adrianna Pinska
Jason van Rooyen
Ashley Rustin
Natalie Simon
51. University of Cape TownOverview for IT Managers – 12 May 2015UCT eResearch
UCT eResearch
Editor's Notes
Data deluge:
Several large-scale multi-national facilities are challenging the way we do science.
The rates of data generation are increasingly massively with LHC requiring 15 PB storage a year. It’s important to realise that these are not acquisition rates but storage rates after post-processing.
At the top of the chart is the SKA which promises to challenge our current technologies which will need to catch up in order to process exabytes of data a day and store 1.5 EB per year.
The paradigm shift is not limited to massive scale infrastructure installations and similar trends are tacking place at the laboratory scale.
Desktop sequencing machines and fast-read-out camera can be purchased from suppliers today which generate in excess of TBs per day.
IT departments have traditionally never had the capacity to support such endeavors and will have to work hand-in-hand with researchers to provide the capacity they need.
In order to better understand PlyC, Monash’s Dr. Shena McGowan (under the direction of Prof. James Whisstock) grew crystals from a PlyC solution, placed them in the Australian synchrotron’s protein crystallography beamline and then bombarded them with x-ray radiation.
Each crystal’s dataset of diffraction images was then transferred to MyTardis (a bioscience research data repository) and archived. Various crystallography techniques and computer programs were used in an attempt to resolve PlyC’s structure. Unfortunately, these were unsuccessful.
In 2010, a new approach enabled Dr. McGowan and Assoc. Prof. Buckle to resolve the structure from the datasets archived in MyTardis.
Resolution of the structure was also assisted by Monash’s high performance computing cluster, which expedited computations.
The structure was published in the Proceedings of the National Academy of Sciences (USA) in July of 2012 and the raw data and images were made available through MyTardis.
What does UCT’s implementation of eResearch look like, what do we do, and how do we accomplish it?
So how do we “create a world-class…”
In this way we are a partner and a broker.
Partner and broker
Talk about grants – advise researchers to come to us early to work out their IT requirements and write it into their grants
What have we done to support the national and institutional research agenda?
Partner from the beginning to develop solutions
What have we done to support the national and institutional research agenda?
Partner from the beginning to develop solutions
“Having a dedicated IT team
just makes economic sense. The
researchers can focus on their core
expertise and let the IT experts take
care of the technology.
The establishment of both HPC
hardware and personnel services at
UCT in 2010 has allowed researchers
to tackle the grand challenge problems
We also engage with facilities which provide technological hubs for researchers
Departments across the university are acquiring technology that is changing the face of research, teaching and collaboration
For instance: In 2014 the Dept of Clinical Laboratory Services acquired a state of the art Olympus VS120 virtual microscopy system.
This effectively digitises old glass slides - so these slide images can now be accessed anywhere with just an internet connection
The Olympus VS120 to be used for research, teaching, by pathologists and Groote Schuur
Centre for imaging analysis: The University of Cape Town plans to establish a world-class visualisation and analysis centre, built largely around electron microscopy.
The eResearch team worked with UCT’s Clinical Research Centre to establish one of the few ‘first in human’ clinical trials on African soil.
We also engage with facilities which provide technological hubs for researchers
Departments across the university are acquiring technology that is changing the face of research, teaching and collaboration
For instance: In 2014 the Dept of Clinical Laboratory Services acquired a state of the art Olympus VS120 virtual microscopy system.
This effectively digitises old glass slides - so these slide images can now be accessed anywhere with just an internet connection
The Olympus VS120 to be used for research, teaching, by pathologists and Groote Schuur
UCT’s Clinical Research Centre to establish one of the few ‘first in human’ clinical trials on African soil.