Capability building and risk management: lessons from Radiata.

SUMMARY

This paper examines the wider lessons to be obtained from the story of the CSIRO-related start-up company, Radiata Inc. It shows how CSIRO's sustained trans-disciplinary capability-building efforts in radio-astronomy helped to produce a generation of electronic engineers well-versed in cutting-edge integrated circuit design and development who were also able to work effectively in commercial environments. Collaboration between radio-astronomers and engineers continued to develop via joint CSIRO-Macquarie University work examining wireless Local Area Network solutions based on mathematical techniques used in radio-astronomy and utilising state-of-the-art chip design methods. This work culminated in the formation of Radiata Inc. and its subsequent acquisition by Cisco Systems in 2001, to be followed by Cisco's withdrawal from wireless chip development in 2004. The paper considers the wider implications of this story, highlighting the importance of trans-disciplinary capability-building to increasing the odds of success in the risky process of innovating. It concludes that CSIRO should continue to develop its options-based approach to valuing R&D outcomes in order to better demonstrate the ways in which capability-building can generate improved odds of success in innovation for a wide range of businesses--provided that they have access to the skilled staff generated by this type of 'rounded' training related to basic research.

KEYWORDS

Radiata; capability-building; innovation; networks; risk; astronomy; semiconductors

**********

This paper considers the role played by CSIRO in the story of the well-known spinoff firm, Radiata. The paper tells a fairly technical story in order to communicate an important message to policy-makers. This message is that long-term capability-building involving close links between blue-sky research and engineering design skills can be extremely valuable. The paper explains how strategic decisions about post-graduate research training and research investments made by radio-astronomers in CSIRO in the 1960s helped to generate some major commercial outcomes some thirty-years later--notably the formation and sale of the company Radiata Inc. to Cisco Systems Inc. for a substantial amount of money.

The story told here covers technical issues because these are critical to understanding what capability building means in practice--beneath the rhetoric that can characterise policy debates over research commercialisation. The paper draws heavily upon a more detailed paper prepared for the Australian Government Department of Education, Science and Training (Matthews and Frater 2003). The names of individuals are intentionally omitted because specific recollections differ between members of the 'community of practice' involved and it may be unfair to name particular individuals in the context of this group-based activity that extended beyond CSIRO per se. The fact that patent infringement litigation is currently taking place reinforces the need for anonymity.

The lesson for policy-making in general, and for CSIRO's future mission in particular, is twofold; first, that this type of long-term capability building generates important option-values as a by-product of blue-sky research and, second, that these options can only be exploited if the necessary engineering and investment risk management skills linked to international networks of 'communities of practice' are also available. The inter-dependence between option values and commercial acumen complicates policy-making designed to encourage 'research commercialisation'. Whilst it is necessary to protect the intellectual property (IP) that arises from publicly-funded research, not least because businesses may appropriate and suppress potentially useful technologies if they threaten the value of corporate balance sheets, there is a significant difference between formal IP protection and extracting significant value from this IP. Consequently, policy-making would benefit from a more nuanced and less prescriptive association between formal IP protection and extracting value from IP.

INNOVATION CAPABILITIES AND COLLABORATION BETWEEN RADIOASTRONOMERS AND ELECTRONIC ENGINEERS

Pushing the scientific research frontier in radio-astronomy involves advancing the technological frontier in radio signal capture and in signal processing. This means stretching the range of frequencies over which signals can be captured and developing methods for distinguishing between signals and background noise. The application of these methods outside radio-astronomy encompasses industrial applications, satellite-based communications, mobile phones and short-distance applications such as wireless Local Area Networks (LANs). These applications require further advances in antenna technology, microchip technology, mathematical applications and other areas.

The close coupling of the technologies used to improve radio-telescope instrumentation capabilities and commercial/defence applications has long been recognised by the radio-astronomy community in Australia and goes beyond simply being receptive to the notion that commercial spin-offs may occur from this R&D work. It extends to a more strategic recognition of the value of building close linkages between the companies that are able to develop and provide critical instrumentation technologies and the R&D, and research training, carried out in order to support radio-astronomy. This results in a system of academic-industry linkages that, from the 1960s, has been far more symbiotic than the simplistic 'linear model' of how basic research translates into industrial applications. The scientists and engineers involved in these capacity-building efforts in blue-sky research recognised that close links with electronics companies could also help to build and sustain a competitive electronics industry in Australia. At that time, radio-astronomy required inter-disciplinary skills in order to design and develop custom microprocessors to handle cutting-edge signal processing challenges and these trans-disciplinary skills had general relevance to industry.

As a result, radio-astronomers have deliberately created an environment for carrying out doctoral research that challenges students to deal with complex projects in which they must deploy leading-edge scientific and engineering skills in order to solve real-world technological problems if they are to carry out blue-sky research. These long-term training strategies did not have particular end-results in mind but sought to create a better capacity to carry out research and innovation in this field.

The core of the story told in this paper is the use of the Fleurs Synthesis Telescope (FST) as a basis for doctoral research training that focused upon real-world radio-astronomy challenges. Several leading figures in Australian (and subsequently global) electronic engineering carried out their doctoral research on the FST within CSIRO. The FST was subsequently taken over by the School of Engineering at the University of Sydney which continued the work. Use of the FST as a focus for 'rounded' doctoral research training had its precursor in the work carried out on the Mills Cross radio-telescope at Hoskinstown in NSW. Following this experience, CSIRO was persuaded to hand over an old field station at Kemp's Creek to the University of Sydney. This became the FST. The FST doctoral experience has produced cohorts of engineers who subsequently went on to work overseas, particularly in Europe and the United States.

CSIRO's Very Large Scale Integration (VLSI) Program was launched when some FST-trained researchers with US experience returned to Australia. The VLSI Program was critical to the development of a key element of the Australia Telescope, known as the Correlator Chip. The Australia Telescope's requirements set a substantial challenge for the VLSI Program because it required putting 100,000 transistors on a chip, at the time an ambitious goal, and making specialised chips for comparing, synchronising and then multiplying signals from different radio-telescope antennae. These chips performed multiplications at 1GHz. The Australia Telescope provided the essential user-focus and first customer for this R&D work and, as a result, enhanced Australia's capacity to design VLSI chips.

In 1984, the CSIRO VLSI Program was spun off into a company called AUSTEK Microsystems. Austek collaborated with CSIRO Radiophysics researchers in an IR&D Board-funded project to produce a 'Fast Fourier Transform' (FFT) chip, subsequently commercialised by AUSTEK Microsystems. This experience at AUSTEK contributed to Radiata's success. AUSTEK also supplied the FFT chip initially used by the bionic ear company, Cochlear, widely regarded as another Australian technology success story.

One other spin-off from applying the FFT chip technology was the formation of Lake DSP in 1991. Lake developed the technology used to produce surround sound in headphones, a technology subsequently licensed to Dolby Laboratories (Lake's first customer) in 1992 and made available on long-haul aircraft following a commercialisation agreement with Dolby Laboratories signed in 1998. Lake DSP continued to work with Dolby, who eventually acquired Lake over the period 2004-5.