WORKSHOP: February 2019

The fourth UKRI Cloud Working Group workshop will be held on 12th February 2019 at the Francis Crick Institute in London. You can see the programme for the day below, and we will be making slides and other resources from the individual sessions available after the event. There is also a crowdsourced write-up via Etherpad that we will be encouraging people to contribute to.

We have a broad range of contributions from across the research community and also good representation from public cloud providers.  This year we focus on containers and workload mobility for our plenary session.   Other sessions focus on mix of application use – from where cloud adoption has reached a mature state – to others where we are examining specific technical and policy related challenges to be addressed.

Timing Plenary / Strand A Sessions
Auditorium 2
Strand B Sessions
Auditorium 1
09:15 Registration + teas/coffee
10:00 Introduction – Philip Kershaw, STFC
10:15 Session 1 – Containers and Workload Mobility
Chair: David Fergusson, University of Edinburgh
Running HPC cloud workloads via serverless functions and Jupyter notebooks – Christopher Woods, University of Bristol
Running HTC & HPC applications opportunistically across private, academic and public clouds – Andrew Lahiff, UKAEA
Cloud agnostic distributed data analysis with Pangeo – Jacob Tomlinson, MetOffice
Question time
11:15 Break – teas/coffee
11:45 Session 2a – Enabling research workloads on public cloud
Chair: Simon Thompson, University of Birmingham
Session 2b – Migrating workloads, hybrid cloud and Federation
Chair: Steve Hindmarsh, The Crick
Increased bandwidth, lower latency and more instance choice for research workloads in AWS – Scott Eberhardt, AWS Cambridge University Data Accelerator and the SKA – John Garbutt, StackHPC
Tailor-Made HPC Storage in the Cloud – Ilias Katsardis, Google Delivering Resilient Access to Globe Climate Projections for Copernicus using Cloud and a Distributed Data Infrastructure – Matt Pryor, NCAS / CEDA STFC
Janet connectivity with Public Cloud – James Blessing, Jisc Case Study of Porting Rfam Pipeline into Cloud – David Yuan, EMBL-EBI
Question time Question time
12:45 Lunch – buffet lunch
13:45 Session 3a Policy Issues
Chair: Martin Hamilton, Jisc
Session 3b – Lightning Talks
Chair: David Salmon, Jisc
 Using Cloud for HPC – Andrew Jones, NAG Bare metal provisioning with OpenStack and containers – Mark Goddard, StackHPC
Assessing Value with HPC-based Cloud – Owen Thomas, Red Oak Consulting MiCADO (Microservice-based Cloud Application-level Dynamic Orchestrator) to support auto-scaling and optimization of cloud applications – Tamas Kiss, University of Westminster
Challenges in making the cloud a strategic part of a University’s research computing infrastructure – Cliff Addison, University of Liverpool Virtual Screening for Drug Discovery using Cloud – Damjan Temelkovski, King’s College London / Francis Crick Institute
Challenges, tools and techniques for migration to cloud – Tony Wildish, EMBL-EBI Jisc Open Research Hub – Tamsin Burland, Jisc
Plugging the Cloud Skills and Knowledge gap – Kenji Takeda, Microsoft
How to completely destroy your cloud HPC project – Cristin Merritt, Alces Flight
 Additional time for Demo pitches
15:00 Break – teas/coffee
15:30 Session 4a – Demos
Chair: Adam Huffman, University of Oxford
Session 4b – International Projects and Collaborations
Chair: Philip Kershaw, STFC
JupyterHub/Pangeo on OpenStack Magnum orchestrated Baremetal Kubernetes Cluster – Bharat Kunwar, StackHPC The European Centre for Medium-Range Weather Forecasts approach to cloud computing – Baudouin Raoult, ECMWF
Horizontal and vertical auto-scaling of containerised applications on MiCADO (Microservices-based Cloud Application-level Dynamic Orchestrator) – Jay DesLauriers, University of Westminster HnSciCloud – Helix Nebular Science Cloud – Joao Fernandes, CERN
EMBL-EBI Cloud Portal for Everyone – David Yuan, EMBL-EBI Moving all our Scientific computing and data to the Cloud, Marc Levesque, Communications Research Centre, Canada (remote presentation)
Reproducible, reliable, reusable analyses with BinderHub on Cloud – Sarah Gibson, Alan Turing Institute The HEPCloud Facility at Fermilab – Steven Timm, FermiLab (remote presentation)
Cluster-as-a-Service on JASMIN – Matt Pryor, NCAS / CEDA STFC Question time
Portable HPC Creation – Will Mayers, Alces Flight
16:55 Final Plenary
Sum-up, feedback, next steps, cloud strategy for research community
17:15 Reception – drinks, light refreshments
18:00 Close