The Active Network

Embedded Linux Strategies

As an old (junior-grade) UNIX™ guru, I'm well aware of the (pre-) history of Linux and the Open Source movement. Spare me a few moments while I describe the first UNIX wave and draw some parallels with the current explosion of interest in Linux. If Windows is your thing, don't turn away! It doesn't matter which side of the fence you stand on, the fence will still be there! ....besides, the grass is always greener on the other side, right?!

Part of the success of UNIX was timing. This portable, cheap, and very coherent OS appeared at just the right time to catch the introduction of the minicomputer. The "mini" meant an order-of-magnitude drop in the cost of computing. It also meant there wasn't enough profit in computer sales to support a proprietary operating system from each vendor. Every vendor's OS development shop was under pressure to match the hardware scale-change with operating software for one-tenth the cost of the prior generation. Further, the customers were suffering out-of-control training costs as more and more relatively inexpensive minicomputers with incompatible operating systems made there way into the organization. A major change was required. A portable, vendor independent OS was just the ticket.

Multics was the first major OS written in a High Level Language. Multics and Algol represented the best the world had to offer in software technology in the early 70's. But Multics development was an expensive and highly academic undertaking. The system itself was much too big and slow for minicomputers and required unique hardware features. These and budgetary reasons led Bell Labs to withdraw from the Multics consortium the meager manpower resources AT&T had offered to aid in development. Miffed, the Bell Labs developers decided to continue their work an an "underground" effort using a PDP-8 mini as the platform. Most of the concepts in UNIX are direct rip-offs from Multics, simplified and streamlined to fit on the minicomputer. Later, rewritten in C, UNIX became "portable" and migrated in the labs from one mini to another. Finally, releasing Version 7 source to universities for next-to-nothing set the stage for the explosive growth of UNIX minicomputers. This was twenty-some years ago.

But part of the success was a near fanatical "religion" that grew up around the idea that UNIX was "open", i.e. not tied to a specific hardware architecture or hardware vendor, and licensed to multiple software vendors. UNIX was viewed as a rebellion that would set all God's programmers free! (After all, we all know the gods smile on fellow creators like us!) Older programmers were feeling the sting of mortality as they saw mainframe vendor after mainframe vendor die. Years of hard work writing proprietary software for now dead hardware left a pretty empty feeling in the stomach, not to mention the now hollow resume! These programmers saw the promise of vendor and hardware independence as a chance for professional freedom and independence. Meanwhile, a generation of programming students cut their teeth on UNIX and provided a ready source of manpower to support the explosive growth of the minicomputer.

Unfortunately, though UNIX was "open" with respect to host architecture, it was owned by AT&T and licensed under fairly legalistic controls. Once AT&T entered the computer market, the "open" nature of UNIX fell into question in the minds of many UNIX system vendors. "Not Invented Here" thinking led to a divergence in features and interfaces between strains of UNIX and the once vaunted "write once, compile to any platform" compatibility ideal fell by the wayside.

There are echoes of these old themes in the Linux and Java movements today. Speaking as a true-believer UNIX-lover who was burned and spurned, I'd have to warn the new generation of believers that the parallels in theme are likely to be followed by parallels in outcome. Human nature has not been repealed. Greed, envy, pride and prejudice can lead to various flavors of Linux and Java, in spite of the "open source" rules of Linux and Sun's "community license" supervision of Java.

But what of the potential of Embedded Linux <>? Is this an order-of-magnitude change of scale similar to the mainframe to minicomputer revolution? Does that provide an opportunity for fueling Linux growth that parallels the history of UNIX on the minicomputer? Certainly the scale change is there. And there is early success in the form of Internet appliances such as NETtel <> and the RAM based "single floppy" Linux. But so is competition in many forms: Windows CE <>, Palm OS <>, and many more proprietary solutions <>.

What could take Linux to the top of the heap?

Impedance matching... ever hear of it? I mention it because this is why Microsoft and Intel have been successful while Multics, UNIX (during the PC revolution), and the Motorola 68000 became "back room" niche players (comparatively), in spite of being technically superior to their competition.

First of all, impedance matching basics:

In it's most familiar usage, it's a ham-radio term that refers to adjusting the impedance of the transmitter to match the impedance of the antenna. If they match, energy flows to the antenna and communications happens. If not, power is reflected to the transmitter and the circuits burn out! I'll leave it to the physicists to fully explain the science behind this phenomenon (hint: it also explains why trumpets have bell shaped ends.... for acoustic impedance matching!) But I'll make a point about Embedded Linux:

There appears to be such a thing as MENTAL "impedance matching". If the idea is easy for entrepreneurs to master and make money using, it flies; if not, the idea crashes and burns (or at least moves into a back room niche.) Embedded Linux may be just the "right" mental impedance to flow smoothly into that market.

DOS, for all it's shortcomings compared to UNIX, fit the requirements of a generation of entrepreneurs dealing with the early PC. They took it and spun it into many things, products that UNIX couldn't have handled, not because UNIX was inferior, but because UNIX was too expensive and too complex at the time. The very same argument applies to the 8086 and the M68000.

This is a fact of the "meta science" of catching technology waves. If a wave in technology resonates with the market opportunities and imagination of entrepreneurs, the force of that wave is magnified out of proportion to any intrinsic technical worth. UNIX missed the PC wave because it had grown to large and expensive during the minicomputer surge. DOS fit the "impedance" of the market and grew to fill the need.

Today's Windows compared to Linux is in the position UNIX was in when the PC came along. Maybe, just maybe, the Embedded Linux opportunity is the one where the Linux community should most seriously focus. Maybe this is the opportunity, the building wave, that Linux fits best.

Can the scenario replay with Linux and the embedded computer market? If the proponents focus on providing just enough technology to meet the need, with just the right emphasis on supporting the entrepreneur, maybe we're in for another big wave!

Copyright © 2000 Information Security Analysis LLC. All Rights Reserved.

Return To The Flink Ink Section


This site is not related to the Microsoft Corporation in any way. Windows and the Windows logo are trademarks of the Microsoft Corporation. ActiveWindows is an independent site. The information and sources here are obtained from series of hard work & research.