BLOG@CACM

Pipelining, Computing Innovation And Economic Growth

By Daniel Reed
February 18, 2009
Comments

Each day brings disturbing news of the deepening economic downturn – declining markets, corporate downsizing and bankruptcy, lost jobs and housing foreclosures, and burgeoning government deficits. We have felt the downturn’s effects on computing as well, with tightening budgets for research, development and education. As everyone looks with a wary eye at discretionary spending, it is instructive to consider the role of computing technology and innovation in our global economy.

The “hockey stick” rule of impact applies, where the function of competitive advantage versus time declines rapidly as a new technology diffuses. The early adopters of new technologies reap the largest benefits relative to their competitors, whether tangible economic ones such as new capabilities and increased efficiency or intangible ones such as market and thought leadership. Later adopters gain the same technical and business advantages but less relative advantage.   To reap the early adopter advantage, of course, one must take commensurately higher risks than the late adopter.

In these perilous economic times, what are the implications for the researchers who produce new computing technologies and those who develop and deploy those technologies? I submit that one must be keenly aware of the lag and hysteresis in the innovation ecosystem. Research investments rarely yield short-term returns. Rather, they create the knowledge base and educated workforce that yield continuing, long-term benefits. They are an investment in the future. (We all have anecdotes of an “overnight success” made possible by a decade of labor.)

This observation leads to two important corollaries.  First, despite our current economic woes, we must continue to invest in long-term, basic research, lest the innovation cupboard lay bare in the future. Second, the computing technologies that will aid our immediate economic recovery already exist – they are the fruits of past research investment. The challenge lies in exploiting them appropriately, both in the public and private sectors.

All of us in computing recognize this fundamental concept; it’s called pipelining, the staged execution of a multi-phase plan. The past pays for the present; the present pays for the future. Let’s make sure our leaders understand the power of pipelining and the consequences of pipeline stalls. Innovation via long-term research investment really matters. Our children will thank us.

-Daniel Reed is Microsoft’s Scalable and Multicore Computing Strategist and a frequent government advisor on science and technology policy. The opinions expressed above are his, not necessarily those of Microsoft or the Federal government. Contact him at reed@microsoft.com or read his other musings at www.hpcdan.org

---

No entries found