Complexity is at the heart of “the software crisis,” a broad term that emerged in the late 1960s to reflect the growing difficulty of developing effective computer programs in the face of ever-increasing computing power. Since then, several schools of thought embracing various approaches to software development have emerged, all with the common goal of coping with runaway hardware capabilities bounded mainly by the curve defined by Moore's law.

Another perspective on this matter is that the enormous and ever-increasing potential of computational hardware is underutilized and will remain so as long as processing capacity growth outpaces the rate at which the software to harness that capacity can be developed. In the late 1950s, the concept of fabricating ICs, called the monolithic idea, was applied to electronics manufacturing to eliminate the runaway cost of manufacturing with discrete electronic components, a problem that became known as “The Tyranny of Numbers.” The software industry seems to be waiting on a similar breakthrough. Yet, when computing elements are placed within their proper context as a subset of electronic hardware, it becomes clear that the development of most electronic systems, especially military systems, is really facing a hardware crisis.

In WWII, the Norden bombsight was a sensitive and protected piece of hardware that was actually removed and guarded after each bombing mission. Bombardiers trained on the device gave an oath to protect its secrets with their lives. This bombsight was essentially a mechanical computer that accounted for several flight variables and precisely controlled the release of the bomb load, enabling accuracies to within 100 feet from altitudes of 21,000 feet. In contrast, many military capabilities are now implemented as customized and upgradable electronics systems installed in mature platforms, such as the C-130, which first flew in 1954 and entered service in 1956, or the F-16, which first flew in 1976 and was fielded in 1979. The removable storage media for these aircraft are now guarded as closely as the Norden bombsight once was.

However, even with this flexibility, the cost of military platforms and even the specialized electronics systems within them have become prohibitive. This has prompted the development of certain military systems as standardized inter-service electronic hardware controlled by customized software and firmware, the software-defined radio for the U.S. Military's JTRS program perhaps being the most significant example.

While this new approach to the hardware of military systems might lead to increased vulnerability of the kind experienced by the Germans in WWII in 1941, when the HMS Bulldog captured a U-boat with its Enigma cipher machine intact, this risk might be reduced by security safeguards and encryption technologies that also fall within the realm of software engineering.

This issue of Defense Electronics will provide designers an opportunity to revisit the balance of software and hardware within military systems as well as in the design of those systems. The systems range from data-centric communications networks, discussed by Stan Schneider of Real-Time Innovations; to advanced UWB radar systems described by Dmitriy Garmatyuk, assistant professor of Electrical and Computer Engineering at Miami University in Oxford, Ohio; and then on to the physical hardware, particularly the antenna requirements, of previously mentioned software-defined radio technology, written by Joseph Jahoda of Astron.

Performance is the real issue in military systems, and hardware often represents a greater bottleneck than software for achieving performance targets. For example, for all the increased sophistication of PCs, you might finish reading this article faster than your machine could boot up because of the latency associated with the hard drive.

Freescale has recently announced the commercial availability of a 4 MB MRAM product with 35 ns read- and write-cycle times. This magneto-resistive RAM technology might eventually replace hard disk technology in certain military applications, as well as in PCs. Hopefully, this technology will also scale with Moore's law, presenting military system designers with new hardware capabilities and at most a relatively harmless firmware crisis.