Summary
print

The SKA Project has a policy of increasing its interaction with industry and, as part of this initiative, a one-day forum for European industry as held at Dwingeloo on Friday 3 December 2004. Organized by the ISPO and ASTRON, the day attracted about 40 people, including about 10 representatives of large trans-national companies.

A notable feature was the video recording of the sessions, an initiative suggested by Harvey Butcher and financed by ASTRON. Selected video excerpts are also available and the wider SKA community is likely to find the afternoon sessions of particular interest.

Goals of the industry day were to present the SKA project to major potential industry partners and stimulate early interest; to highlight European SKA activity; to look for examples of best-practice in industry-science links; to air the issues surrounding intellectual property and other legal issues; and to seek input to the development of a strategy for SKA-industry liaison.

Companies represented at the gathering included Alcatel, Cisco, HP, IBM, MAN and Philips. The day was framed as an informal briefing and discussion forum, with the ISPO giving initial disclaimers relating to the forum’s link to any future commercial liaisons. Current direction in the SKA project were described by a number of speakers including Harvey Butcher, Richard Schilizzi, Peter Hall, Arnold van Ardenne, Jan-Geralt bij de Vaate, Parbhu Patel and Peter Wilkinson. After presentations by the SKA speakers, company representatives were invited to give their views on the project, including their perspectives on how it may interest the industry sector(s) in which they operate. Michel Apers (Program Manager, Business Unit Observation and Sciences, Alcatel Space), Ruud Haring (Team Leader, Blue Gene Chip Design, IBM Research) and Brett Biddington (Space Initiatives Manager, Global Space Team, Cisco Systems) gave brief presentations, all of which raised interesting technical and logistical issues to be faced by the SKA project.

An advantage of holding the Industry Day at ASTRON was the readily-available experience of those actually building a new-generation telescope, namely LOFAR. Eugene de Geus (Director, External Affairs, ASTRON) did the SKA project a great service by outlining LOFAR experiences to date, including experience flowing from R&D programs which are tightly integrated with industry. Eugene also chaired an interesting afternoon panel discussion which focused on logistical and legal issues in multi-national projects. Panel members Niels Eldering (ESA Technology Transfer and Promotion Office), Stephen Kahn (ProContra bv) and Jan Reitsma (LOFAR Project Manager) provoked a spirited discussion which extended well into an informal drinks function held after the close of the sessions.

Logistical and financial support for the Industry Day came from the ISPO and ASTRON. In particular, Astrid Marx and Stephanie Voegele provided valuable behind-the-scenes assistance.

A few specifics to emerge from the day include:

· Advances in high performance computing will make it increasingly attractive to consider HPC as at least a partial replacement for “traditional” DSP techniques. The processing rates demanded by SKA are easily consistent with predicted peak computation speeds and there will be substantial gains in the average speeds of clusters. HPC gains beyond ~2010 will no longer come from device scaling, nor will they come easily from any other single path. Instead, speed gains will arise from a succession of one-off “tricks”, system-level (not just chip) optimization, and advances in high-speed I/O technology (e.g. active equalization). Power consumption (dissipation) and reliability will be issues, but probably resolvable ones.

· The massive deployment of networked RF identification systems and other wireless sensors will dramatically affect low-cost RF systems and network design, leading to much greater development in “edge-server” architectures: architectures in which large amounts of pre-processing are done at the edge of a network which ultimately feeds data to hubs. Some of this development may be applicable in very large distributed systems such as the SKA, either in the context of system monitor and control (based around new-generation internet protocol standards) or, more interestingly, for signal or data processing.

· The SKA is a big enough project to influence substantially the national networking strategies of some countries. Dual scientific and commercial use of some SKA communications infrastructure is an obvious route but the project influence extends to e.g. impacting government and market views of alternative carriers, such as satellite and military operators.

· There are substantial challenges in meeting international legal requirements for a project needing industry involvement in both pre-competitive and procurement phases. Without care, and a diligent arm of management devoted to the challenges, it is possible to lock-out early collaborators from participation in later procurement activities. Rules are similar in principle in major blocs such as the European Union and the USA but there are differences in detail. Nevertheless, there are examples of successful collaboration: the European contribution to HST was mentioned as one of them. Framing the SKA project to make possible observance of, for example, World Trade Organization rules, could dictate the way work packages are allocated and performed and, effectively, the structure of the project. It is even possible that the project structure, incorporating industry participation in various phases, will influence the system architecture: the idea of defining some system blocks in terms of what can be developed and procured via an interested industry partner is one pioneered by LOFAR.

· Several industry attendees expressed a view to the effect that, to meet advertised milestones with industry involvement or even vocal support, the SKA project office now requires a strong industry and international legal branch.

· On a lighter note, the audience was grateful to Brett Biddington for adding three important overlays to the usual seven networking layers. Sitting on top of the applications layer was religion (“that’s how we’ve always done it”), politics (“its my data!”) and economics (“we can’t afford it”). Those present resolved be wary of these mind-sets when planning the SKA.

The European industry day was undoubtedly successful in airing and raising many issues for the SKA project. Concrete technical collaborations with national consortia offer the best short-term possibilities for industry involvement. In addition, industry participation in EWG and other international Task Forces is encouraged, and forums such as that held at Dwingeloo are useful in highlighting these opportunities. The next stage is for the SKA project to form an additional EWG sub-group, to be known as the Industry Liaison Task Force and chaired by the International Project Engineer, with the role of advising the project on industry matters. The ILTF will generate a second whitepaper on industry policy by the end of 2005, extending the commentary of SKA Memo 52 beyond pre-competitive alliances. The suggestion from those present in Dwingeloo to expand the industry linkage role of the ISPO were well-taken; this of course is a challenge to a project still seeking funding for technical and scientific management but it may, nonetheless, be essential if engineering and funding milestones are to be met.

Finally, it is clear that there is a place for future SKA-industry forums, including gatherings to address the legal and trade issues which have now been partially exposed. Several attendees from trans-national companies expressed interest in a North American forum and this suggestion will be followed up by the ISPO.

print

More
Program - Presentations - Pictures - Participants