Technology Library

Sep 1999
Engineering Systems on a Chip: The Bloody Revolution in Tools, Methodologies and Power

Abstract

Systems engineering is not a quiet backwater. It is an arena of political and economic realignment and foment. And like any revolution in-progress it is the harbinger of great and unpredictable change bringing with it huge opportunities hand-in-hand with equally impressive threats. Nothing less than corporate power and bureaucratic structures are being reordered by the new realities of engineering systems on silicon. The 30 years of hardware dominance in silicon electronic engineering is being torn apart by the recognition that the complexity of modern systems is determined by its multi-functionality, adaptability and flexibility – attributes that, in an economic sense, are best realized in software. 

The genesis of this revolution has been the stunning success of the silicon engineers, which ironically, as with the yin-yang cycle, has carried with it the seeds of its own diminishment, at least for the half-turn of the next cycle. As silicon technological progress marches through 0.18m minimum feature sizes to 0.15m to 0.13m in the next couple of years, and then to sub 0.10m, the number of transistors on a chip will approach, then exceed, one billion. The majority of these transistors will be consumed as on chip memory devices. Memory is most useful in programmed devices and with processors executing in excess of a billion instructions per second when implemented in 0.1m silicon technology, programmed devices will progressively phase out many special purpose hardware devices.  

The dismantling of the power base in the semiconductor divisions of the major electronics companies to accommodate software engineering and systems integration on an equal footing with hardware engineering is a battle in full flight. Like the Iliad, where Greek’s fiercely and intermittently battled Trojans for ten long years, the outcome of the systems engineering battle is certain. Unlike the complete destruction of the city of Troy, the resulting realignment will be a triumvirate between – hardware, software and mechanical designers. The sooner the battle is won the sooner the potential of the three powerful potentates will yield novel systems and architectures to dazzle the techno-enervated masses. Like all power-sharing structures, it is unstable but ultimately governable by the dour and pragmatic economics of survival. Already fleet-footed start-up companies are demonstrating the fecundity of the new godhead – carpe diem!