Embedded Linux is revolutionizing the embedded systems software market. The advantages Linux offers relative to traditional proprietary embedded software solutions are tremendous. These advantages translate to significant savings in total cost of ownership and time to market for embedded developers and OEMs, and increased ability to highly differentiate products. The result is a dramatic shift to Linux as the strategic embedded operating system platform of choice in industries of all types. This shift is occurring today and is accelerating as the technology sector initiates a new business growth cycle.

The competitive advantages of embedded Linux are so significant that even companies satisfied with their proprietary solutions will be required to make this shift to remain competitive. The result will be the end of the traditional fragmentation of the embedded operating system (OS) industry. Embedded Linux will grow to be the dominant embedded OS solution with a majority market share.

The Embedded Market and the Embedded System Software Crisis

The embedded systems market can be defined very broadly as product teams executing a development lifecycle, who must address integrated computer system hardware and software requirements. Alternatively, the embedded systems market can be defined from a product perspective as products that internally incorporate computing technology, but which are not marketed as traditional desktop or server computers.

Such definitions include a plethora of product types, including telecommunications servers, routers, process control systems, factory automation equipment, military and aerospace systems, medical equipment, test and measurement devices, in-car information and entertainment systems, set-top boxes, and personal video recorders, cameras, phones, and PDAs. This list is expanding to include non-traditional product types, as computing intelligence and connectivity become pervasive in home appliances, wristwatches, vending machines, toys, etc. Even running shoes may soon be an embedded system product category!

There are two dimensions of rapid expansion of the embedded system software market: the first is the growth of product categories, and the second is the growth of requirements for the system software in the embedded device itself. For example, a toy manufacturer that previously dealt with very limited micro-controllers and small amounts of firmware may need to embed a complete OS environment with multi-programming, TCP/IP connectivity, and Bluetooth wireless I/O. Similarly, a multi-national consumer electronics company might find its product teams faced with making decisions about the purchase and incorporation of millions of lines of OS software into their product lines, as every product now requires multiprogramming and Internet connectivity.

With such an explosion of internal costs and of the complexity of product design and development, there is an impending risk of a software crisis throughout technology-based product industries.

The Unique Values of Linux for the Embedded Market

Against this backdrop of powerful forces expanding demand in the embedded marketplace, embedded Linux has arisen as a new and very different alternative. An embedded Linux solution targeted toward product engineering teams has a number of unique values.

First, embedded Linux offers a simplicity and convenience which is unparalleled by traditional embedded system software products. The simplicity starts with “it's just a UNIX target,” which leverages the UNIX development skills that so many software engineers possess. Even more convenient, the universe of software engineers is increasingly Linux-literate and is therefore immediately comfortable with the core development tools and the programming paradigm for Linux embedded system software. Compared with the difficulty of learning, structuring and becoming productive in a proprietary tools environment (e.g. Wind River's Tornado), embedded Linux is vastly more simple, familiar and convenient.

Additionally, there are several significant benefits which derive from using Open Source Software which are, by definition, impossible to replicate with proprietary or in-house-developed embedded system software. These include . . .

• Vendor independence — Multiple vendors offer embedded Linux product solutions with support, training and professional service options, using a variety of business models with a variety of methods of differentiation. Additionally, the fundamental source technology is available, making self-development and self-support of a solution a viable (albeit expensive) alternative. Traditional exclusive dependence on a single vendor, and the risk contained therein (witness Wind River's absorption of pSOS and Mentor Graphics' of VRTX and Nucleus), is eliminated.
• Availability of trained expertise — Linux-savvy engineers are available throughout the world. Linux is the foundation for systems education in universities because of its openness. Books on Linux usage, programming and internals number in the hundreds.
• Ability to create optimal designs through internal customization — This tremendous benefit enables types of product differentiation not available at reasonable cost with proprietary products.
• Early availability of hardware support — Linux is by far the preferred bring-up and initial support vehicle for hardware manufacturers. This hardware support starts at the microprocessor level, where Linux availability outstrips any other OS in the history of computing, and continues to the I/O level and beyond. Additionally, the vendor independence assures that if one Linux vendor is unable or unwilling to invest in the needed hardware support, there are alternative solutions.
• Early availability of new technologies in general — Linux is also the preferred vehicle for computing system hardware and software innovation. Cutting-edge technology will almost always be available for Linux ahead of other system software platforms, and certainly relative to proprietary system software.
• Lower cost — Linux starts by providing a royalty-free deployment platform. Next, Linux deployment reduces costs to hire and train engineers, enhances the ability to customize, and offers a simpler and more convenient development environment. The result is a vastly reduced total cost of ownership. The availability of commercial distributions that provide a complete starting point for embedded system development only serves to further reduce costs relative to cobbling together and maintaining an embedded Linux development effort in-house.

Embedded Linux as the strategic choice

It is insightful to consider these factors from the perspective of a CTO or VP of engineering of a technology products company with growing embedded computer system requirements. Faced with the need for dramatic expansion and/or upgrade of embedded system software within his or her company, what is the likely thought process of such an executive?

• Our products have increasingly complex system software requirements, including Internet connectivity, a need for the most competitive leading-edge microprocessor technology, and a need to support rapidly evolving and extremely complex I/O technologies.
• Our needs for such system software cut across a large number of different product organizations and teams, and are moving into product categories that previously had no computing content at all.
• My competitive situation doesn't tolerate a high cost structure for system software. Current budget constraints will not support the in-house teams formerly maintained; engineering resources must focus on our own value-added technology to remain competitive. Also, the royalty component of our current off-the-shelf system software in our cost of goods is large and impacts our margins.
• Selection of a common, strategic system software platform will help the company avoid getting mired in a multiplicity of solutions, each with a high cost structure and no overall leverage. We know from past experience that disparate platforms inflate costs and increase product development cycles.
• Proprietary or open source? Proprietary will . . .
  • Lock our company into one strategic vendor; if that vendor fails to serve us well, we suffer a long-term competitive disadvantage, or must switch platforms again with a loss in time to market and an incursion of accompanying costs.
    
  • Limit our ability to adapt technology as needed to maximize competitive advantage.
    
  • Limit and/or delay our options on hardware selection.
    
  • Increase cost of goods through per-unit run-time royalties.

• Conversely, selecting an open source Linux platform will . . .
  • Allow selection of a single, strategic vendor that can provide a ready-to-use development solution with full support and training.
    
  • Allow us to change suppliers over time at relatively low cost if a vendor's integrated product or service lags behind the offerings from other vendors.
    
  • Allow us to perform the necessary amount of system platform customization needed to maximize our competitive differentiation.
    
  • Enable us to choose the best available computing hardware, free of constraints imposed by our own system software platform.
    
  • Allow us to utilize the most current and capable I/O in our products.
    
  • Help us leverage the most current innovations in software technology in our products.
    
  • Foster development of our own internal competence to enable us to build highly differentiated product value.
    
  • Permit us to take advantage of a worldwide supply of professional service expertise focused on system software technology.
    
  • Allow us to recruit from a global pool of engineers trained in the technology.
    The conclusion that embedded Linux is the best strategic choice is inescapable. The unique values of embedded Linux provide tremendous business value, everything from reduced costs, and reduced time to market to increased competitiveness, for technology companies designing and developing products with embedded computing technology. These benefits are driving a massive shift in multiple industries to embedded Linux today.

Embedded Linux as Disruptive Technology

Linux as an embedded operating system solution for product development teams has all the properties of a “disruptive technology” as defined by Christensen in his book, The Innovator's Dilemma. Disruptive technologies:

• Are simpler and more convenient to use than the current dominant technologies.
• Provide new and different value to particular market segments or product categories that are not provided by the current dominant technologies. Frequently, such a market segment or category of product is non-existent prior to the emergence of the enabling disruptive technology.
• Are typically lower in cost than existing dominant technologies.
• Frequently under-perform relative to traditional dimensions of performance or value that have been the basis of competition for the current, dominant technologies.

Embedded Linux clearly demonstrates the first three properties above. With respect to under-performance, embedded Linux has frequently been maligned by proprietary OS vendors as too slow, too large, and lacking tools. These are the traditional dimensions of competitive differentiation in the proprietary embedded OS market. Early adopters have found the unique values of embedded Linux surmount these perceived shortcomings. Over time, investment by MontaVista Software and others in open source solutions to these traditional competitive needs has brought Linux up to and over the minimal requirements of the mainstream market, and in some cases to a strongly competitive position, in exactly the manner that Christensen outlines in his theory of disruptive technology.

It is interesting to note that the values of Linux for embedded product design teams are not values that apply to Linux for the desktop and general server markets. Linux for desktops and servers is not simpler nor is it more convenient than the Microsoft product alternatives. The open source derived values for engineering teams don't apply to the desktop and server users (who are not development engineers). In Christensen's terms, Linux for desktops and servers is a sustaining technology, competing almost exclusively on cost. In the server market, the cost benefit has enabled Linux to penetrate very deeply in the market. However, it is only for the embedded market (specifically for embedded system engineering teams) that Linux becomes a truly disruptive technology with unique values that drive a revolutionary shift.

The values of embedded Linux are also completely independent of the system(s) of choice for the desktop and server environment in a product engineering organization. An existing major investment in a Sun Solaris or Microsoft Windows-based software engineering environment need not be jettisoned in favor of desktop or server Linux to take advantage of the values of embedded Linux. Complete embedded Linux cross development toolkits hosted on Solaris, Windows, and Linux are available today from MontaVista Software and several other commercial embedded Linux toolkit suppliers.

Summary: The Innovator's Advantage

Although it is not yet entirely visible, due to the confidentiality surrounding many development projects, the Linux embedded system software revolution is not a future event — it is underway now, and moving at full steam! Consider, for example, that MontaVista Software alone has nearly 500 active customers for its professional embedded Linux product solutions — with each such customer representing revenues of tens to hundreds of thousands of dollars, and some making million-dollar commitments. These customers represent every conceivable product segment and include numerous brand name Fortune 100 companies. Additionally, many are in the process of adopting embedded Linux as their strategic embedded system software.

We are thus rapidly approaching the “early majority phase” of the development of this new market (as defined by Geoffrey Moore in Crossing the Chasm). The resurgence of technology investment and spending in the coming months will drive embedded Linux into an exponential growth phase. The traditional fragmentation of the embedded system software industry will be replaced by an overwhelming dominance of embedded Linux as half, and then more than three-quarters, of new product development programs move to Linux by 2005. The combination of a thoughtful analysis of the disruptive values of Linux for embedded product developers, plus observation of the breadth and depth of embedded Linux business successes despite an extremely challenging technology business cycle, leads to no other conclusion.

Why are technology companies buying a professional embedded Linux product solution, rather than putting together and maintaining one of their own? After all, “it's just open source,” and “Linux is free.” The answer is simple. They are choosing to buy, not build, because purchasing a professional embedded Linux product solution saves substantial time and money over do-it-yourself options. The buy-versus-build decision becomes simple when the cost and complexity facts are considered.

Consider, for example, the complexity of a well supported professional embedded Linux product solution such as MontaVista Linux 2.1. It is composed of 29 million lines of source code, drawn from 19 different unsynchronized and non-integrated code repositories, and it took nine calendar months and approximately 17 engineer-years to construct, qualify, and deliver. Additionally, MontaVista Linux 2.1 targets over 70 board computers, 24 microprocessor architectures, and 11 different host computing environments out of the box, and ships complete with ready-to-run binaries, bundled unlimited technical support, and defect corrections and product update access.

Today, companies must focus their development resources on their core competencies and core differentiated values. Replicating this level of effort for a component technology would be a costly error. Only rarely will a company need to innovate substantially within Linux system software, justifying financially a “build” approach to embedded Linux, just as few companies today can financially justify designing their own processors or I/O components.

Christensen wrote his book and took his title from the perspective of traditional product companies struggling with innovative technology, under attack by innovators aggressively bringing disruptive technologies to market. By aligning their strategies and business models around the revolutionary and disruptive nature of Linux for embedded system developers, professional embedded Linux product solution providers are driving the ongoing revolution in embedded system software, fueling the fire of embedded Linux's disruptive value, and closing the performance gap between Linux and traditional proprietary solutions.

What is a dilemma for proprietary platform vendors is The Innovator's Advantage for the companies who are driving the ongoing embedded Linux revolution.

---

About the author: Kevin Morgan is Vice President of Engineering at MontaVista Software. He has 20 years of experience developing embedded and real-time computer systems for Hewlett-Packard Co. Experienced in operating systems and development, Kevin was a member of the HP 1000 computer software design team. While at Hewlett-Packard, he worked as an engineer, project manager and section manager spanning the development of five operating systems. As HP-UX Operating System Laboratory Manager, Kevin was responsible for overall HP-UX release planning, execution and delivery for Hewlett-Packard server computers. Kevin has been leading the MontaVista Software engineering team since joining the company in March 1999. Kevin obtained his B.S in Computer Science from the University of California, Santa Barbara and earned his M.S in Computer Science at the University of California, Berkeley.

---

---

 
This article was originally published on LinuxDevices.com and has been donated to the open source community by QuinStreet Inc. Please visit LinuxToday.com for up-to-date news and articles about Linux and open source.