NeIC Conference 2013: Report on "Opening Session"
"Welcome" (Petter Kongshaug)
- High quality networks and federated access needed
- Tools and services (HPC, storage, visualization) needed for collaboration between research partners
- Have reached a "turning point" in HPC. Misc. tools are getting developed, access easier, etc
- NeIC is working closely with the local organizations such as SNIC.
- Important to sync activities in the Nordics
- This conference should be the starting point for the next level of Nordic collaboration
"Official opening" (Pål Sørgaard)
- Moving towards Exaflops
- We can most likely meet the researchers needs based on the current pace in computer development
- We are in a field where things are changing rapidly - the pendlum swings back and forth
- A game change: From sequential devices (mainframes) to grid, cloud and back to centralized architectures
- Nordic countries always played a central role in the development
- Advanced networks are in place
- Mission to collaborate and gather all competence in the Nordic area
- Sharing of data often very complicated and difficult because different rules and challenges. Trust important
"Welcome from NTNU" (Torbjørn Digernes)
- Information technology is an important and great driver
- The development is definitely not slowing down, instead it's accelerating
- What use to be "traditional engineering" have changed with the introduction and increase in computational power.
- If you don't have access to information at the rate needed you will loose your pace compared to others
- Vulnerability: We are getting increasingly dependent on information technology. If the systems break our work is immediately affected.
- Computer systems have probably the shortest lifespan compared to other technical equipment. Around 5Y (from mfg to museum)
- Updating the infrastructure in a clever way over time is a huge challenge
"A Vision for Nordic e-Infrastructure Collaboration" (Gudmund Høst)
- An IT infrastructure is a complicated subject that is getting more and more complex over time
- Hardware and software changes rapidly
- The human asset (the people, the experts) are the sustainable asset that are most important
- How measure success? Research results (such as the Higgs particle), Personal success (such as new positions)
- NeIC is not a funding agency. It's however happy to co-found.
- NeIC is a framework for collaboration that will orchestrate the collaborative work between research groups and countries in the Nordics
- No strict rules apply regarding how the collaboration is performed between researchers or countries
- NeIC have achieved a sustained will/support for 10Y from the Nordic countries
- NeIC would appreciate to get input regarding improved ways for Nordic collaboration and new project ideas.
- Most motivating is to provide solutions that will add value to the researchers
Nordic comparison:
- Sweden: slim layer of coordination, SNIC is a small organisation, 6 centers
- Finland: CSC do both coordination & service
- Norway: Similar to Sweden, slim coordination layer, 4 centers do the service
- Iceland: As Finland but a more slimmed organization
- DK:
"Enabling excellent science through High-Performance Computing" (Kenneth Ruud)
- HPC speeds up the scientific discovery process
- Computations can provide details that are not experimentally observable, allowing the understanding of chemical processes and properties to be refined.
- Certain research wouldn't be possible without HPC assistance
- HPC is key for science
- Industrial examples: Manufacturing: Shortening the lead time, fewer physical prototypes. Oils & Gas: Finding the well with high precision saves time and money.
Some trends in HPC challenges:
- Multiscale problems: symbiosis
- Ensemble averages: real life problems have no clear initial conditions
- Exascale challenges: more than one roadmap, better integration of hw & sw technologies
- BIG data challenge: How to store it? How to analyze it? Very rapidly increase in data size.
EU level:
- Europe should continue to provide world leading HPC infrastructures
- Leadership and management of HPC infrastructure at European level should be a partnership between providers and users
- New challenges require long term committments. A planning horizon of 10-20Y and a rolling 5Y upgrade plan is needed
- Urgent need for development of improved e-Science algorithms and software
- Training programs at different levels are important
Nordic level:
- Can we create Nordic excellence in science through Nordic collaboration in HPC?
Key questions:
- What is needed between EU and the Nordic level?
- Is it financial sustainable to have multiple levels?
- Is the national level needed?
Panel Discussion
Session Summary
- There is a high degree of trust in the Nordics, for example for government and for research.
- NeIC has strong political support.
- Data sharing, and open access are being pushed by the Norwegian government.
- People are the key part of e-Infrastructure. Hardware & software are ephemeral and/or disposable (5yr life time). People are sustainable.
Lessons Learned
- As scopes of scientific challenges expand, the need for organising international collaborations increases.
- Trust is key for information security, and e-Infrastructure must work to maintain and keep earning that trust.
- Can we create Nordic excellence?
- Is there a gap between PRACE, Nordic, national levels? Should we keep on funding all of these levels?
- Can software and hardware co-design help productivity?
- Can Nordic collaboration increase funding?
- Nordic perspective will enable e-Infrastructure engagement in new areas.
Future Directions
- NeIC and NORDUNet must add value nationally, and represent the Nordics in a larger context.
- NeIC is framework for collaboration, not a company.