FP7 Case Study: NanoStreams, Dimitrios S. Nikolopoulos, Queen's University Belfast

494 views

Published on

Presentations from Horizon 2020 Information Seminar held at Hilton Hotel 28th November 2013

Published in: Technology, Business
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
494
On SlideShare
0
From Embeds
0
Number of Embeds
2
Actions
Shares
0
Downloads
4
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

FP7 Case Study: NanoStreams, Dimitrios S. Nikolopoulos, Queen's University Belfast

  1. 1. School of Electronics, Electrical Engineering and Computer Science FP7 Case Study: NanoStreams Professor Dimitrios S. Nikolopoulos
  2. 2. School of Electronics, Electrical Engineering and Computer Science About me • Professor and DR, School of EEECS QUB • US academic background • Computing systems researcher – High-performance computing – Processor and memory architectures – System software and firmware – Languages and compilers – System integration November 28, 2013 H2020 ICT Info Day 2
  3. 3. School of Electronics, Electrical Engineering and Computer Science How did NanoStreams start? • Neither from an EU ICT Work Programme nor from a call target – A clearly identified challenge that merits investigation from an excellent consortium • Critical mass in the financial trading sector in Belfast – Extreme performance requirements, at odds with technology trends towards energy-efficient servers • Solutions considered required disruptive changes that span the entire computing software & hardware stack – Could not be possibly solved by a University, SME, or even multi-national – First hint to consider an EU project November 28, 2013 H2020 ICT Info Day 3
  4. 4. School of Electronics, Electrical Engineering and Computer Science How did NanoStreams evolve? Context – Real-time analytics on streaming data require extremely high performance and low latency – Critical for important markets: capital markets, ICU monitoring, fraud detection, business intelligence,… – Currently supported by servers with high TCO • Idea – Can a low-cost server of a size of a wallet be used instead? • Motivation – Facebook stripped-down micro-servers • November 28, 2013 H2020 ICT Info Day 4
  5. 5. School of Electronics, Electrical Engineering and Computer Science How did NanoStreams evolve into an EU ICT project • Contemplated several funding alternatives – Hard to involve SMEs and industry as funded partners in RCUK – Limited funding, small “demos” supported from instruments like TSB • Critical mass of expertise available in the UK but important components missing: • Compilers for application-specific processors • Real-time operating systems • Major server vendors not UK-based November 28, 2013 H2020 ICT Info Day 5
  6. 6. School of Electronics, Electrical Engineering and Computer Science Reading (behind) the EU call “Synergies between High Performance Computing and Embedded Systems to address a common challenge” • High performance from a computing system with a low carbon and space footprint – What is common between a smartphone and a datacentre server box – Embedded many-core processors running highperformance parallel software November 28, 2013 H2020 ICT Info Day 6
  7. 7. School of Electronics, Electrical Engineering and Computer Science NanoStreams (in a picture worth a thousand words) November 28, 2013 H2020 ICT Info Day 7
  8. 8. School of Electronics, Electrical Engineering and Computer Science NanoStreams Consortium November 28, 2013 H2020 ICT Info Day 8
  9. 9. School of Electronics, Electrical Engineering and Computer Science Why was NanoStreams funded? • Clearly identified challenge and a need • Perfect match to the call • Strong participation by industry and in particular SMEs – SMEs that develop (and not just use) technology – Large corporates that are ideal customers for SMEs – Job creation, economic impact • Internationally excellent consortium, complementarity • A unique product proposition based on European technology – With potential US uptakers November 28, 2013 H2020 ICT Info Day 9
  10. 10. School of Electronics, Electrical Engineering and Computer Science Importance of networking • Instruments leveraged to establish a consortium – ECIT/CSIT commercial team – European Network of Excellence in High Performance and Embedded Architectures and Compilers (HiPEAC) – Partnership for Advanced Computing in Europe (PRACE) – Personal contacts – “Dear John” emails November 28, 2013 H2020 ICT Info Day 10
  11. 11. School of Electronics, Electrical Engineering and Computer Science Proposal writing process • 10-page draft from Coordinator circulated September’12 • Work-packages and roles discussed in F2F meeting in Paris and drafted October’12 • Distributed writing – Each WP lead by a different partner – Part B1 written by technical coordinator, with on-demand contributions by other partners – Part B2 written by a research engineer with prior experience in coordinating EU projects – Part B3 written jointly by one SME and one corporate • First draft completed November’12 • Continuous major and then minor revisions until submission in January’13 November 28, 2013 H2020 ICT Info Day 11
  12. 12. School of Electronics, Electrical Engineering and Computer Science The importance of chemistry • Partners have previously worked together in clusters – QUB, IBM, FORTH (TEXT project) – Analytics Engines, ACE (commercialisation effort) – Neueda, Credit Suisse (commercialisation effort) • Excellent inter-personal relationships • Understanding the cultural background, business environment and financial climate was important November 28, 2013 H2020 ICT Info Day 12
  13. 13. School of Electronics, Electrical Engineering and Computer Science Working with the Commission • Get to know your Project Officer as best as you can – Technical background, cultural background, preferred means of communication, style, track record, experiences of others… • Experience may vary dramatically – NanoStreams: 30-minute negotiations meeting, 4 talking points, 1 major and 3 minor revisions of the DoW – CACTOS: 3-hour negotiations meeting, including a 1-hour project presentation 37 talking points, ~50 revisions of the DoW. 10+ major revisions – ASAP: day-long negotiations meeting, 127 talking points (12 pages of comments), 130+ revisions of the DoW November 28, 2013 H2020 ICT Info Day 13
  14. 14. School of Electronics, Electrical Engineering and Computer Science Working with the Commission • Your PO may change any time – Unnamed FP7-ICT-Call 8 project – 4 POs in 3 years – PO A, B, and C conducted project reviews and approved Good and then Excellent ratings – PO D returned all dissemination deliverables as unacceptable, threatened with a poor rating, and then threatened to recover the funding of a major UK semi-conductor IP company November 28, 2013 H2020 ICT Info Day 14
  15. 15. School of Electronics, Electrical Engineering and Computer Science The benefits of participating… • A broader talent base – Increased chances to work with the very best in a number of disciplines • A broader funding base – More funding instruments, higher budgets • A broader market base – Industry engagement, industrial R&D and take-up are essential (and not cosmetic) requirements of these projects • More pathways to impact – Companies working with academics to develop new products and services November 28, 2013 H2020 ICT Info Day 15
  16. 16. School of Electronics, Electrical Engineering and Computer Science The benefits of not participating… • Administrative overhead – Extensive reporting requirements – Annual (and oftentimes semi-annual) reviews – Extensive travel • Management difficulties – Integration towards a common vision, common set of tools, common outcomes – Multiple accounting systems, research cultures, product development cultures, people cultures… – Under-performing partners, weakest links, workload imbalance… • Inappropriate instrument for blue sky research – Projects and not programmes – Tight timeline, small horizon, many checkpoints – Too much integration effort November 28, 2013 H2020 ICT Info Day 16
  17. 17. School of Electronics, Electrical Engineering and Computer Science Would I do it again? Absolutely! but never attempt to coordinate more than one projects simultaneously 
  18. 18. School of Electronics, Electrical Engineering and Computer Science FP7 Case Study: NanoStreams Professor Dimitrios S. Nikolopoulos

×