Embedded Internet and ZiLOG - a White Paper

Bill Giovino

Introduction - the Evolution of Embedded Systems and the

Of all the semiconductor industries, the Embedded Systems marketplace is the most

conservative. Engineering decisions in this market are usually conservative, leaning

towards established, low-risk solutions. Because of this, the basic infrastructure of

embedded systems has only evolved slowly over the past ten years; 8-bit

microcontrollers are still the workhorses of the industry, with 16-bits slowly gaining

ground. C Compilers, debuggers, and RTOSes gradually took root in this industry

over the past ten-years. Emulators have slowly given way to high-speed software

debugging techniques. Networking of embedded systems has gradually become

more prevalent than stand-alone systems.

The keywords here are ‘slowly’ and ‘gradually’. Embedded developers are a

conservative lot and are reluctant to jump onto a new technology where there is no

clear advantage in project development. That is why well-publicized technologies

such as fuzzy logic and embedded java have seen so little market acceptance –

these technologies were not attempting to solve any existing development problems.

While the landscape of embedded systems is vast and encompasses a great deal of

application diversity, more and more embedded systems are now interconnected thru

serially networked systems. Many high-volume embedded systems applications are

physically networked via UARTs, CAN, SPI, or Ethernet. Supporting these various

hardware standards are software standards, many of them voluntary, conformity of

which is subject to change based upon the whim of the design engineer. Because of

this, two UART-based systems, for example, probably will not be able to network

between each other even though they are based on the same hardware standard.

In addition, many conventional embedded systems are primarily used in control-

oriented applications in sensing and affecting external events. Many of these

microcontrollers are managed by human controllers by some sort of man machine

interface (MMI) – for example, a cash register, a cell phone, a T.V. screen, or a PC

interface.

It is this MMI that often represents the most costly investment in the system’s

development in both time and money.