SFF market: Healthy diversity, or option overload?

Editor’s note: It’s a question that hedonists and chocolate lovers likely spurn: Can you ever have too much of a good thing? In an embedded Small Form Factor (SFF) industry flooded with dozens of standard and proprietary module sizes, some might answer in the affirmative, asserting that the large quantity of module options fragments the market and makes design wins difficult to achieve. Others might disagree, stating that the vast number of SFFs fuels competition and thus innovation, providing a boon for both vendors and customers. To get a pulse on the different perspectives in the marketplace, we posed several questions to executives from leading SFF companies and assembled excerpts from their responses in this virtual roundtable discussion.

21
Figure 1: Executive panel
(Click graphic to zoom by 1.9x)

SFF: Do you think the SFF market is fragmented, with a surplus of standards and sizes that make it difficult to gain a significant market share? Why or why not?

 

BURCKLE: Yes, I think the market is fragmented with quite a number of SFF boards from different vendors. But no, I don’t think there is a surplus of standards since most of these boards are developed and supported by a single company with no independent industry standards group behind them such as SFF-SIG, PICMG, VITA, or the PC/104 Consortium.

The great number of form factors makes it difficult for designers to wade through the process of locating, evaluating, and selecting a specific form factor board for their projects. Yet if they consider a truly standard SFF, then their task is narrowed considerably. It then becomes a matter of comparing specifications, I/O expansion options, temperature range and other harsh/rugged operational issues, CPUs supported, enclosures, ease of developing any custom boards, and multisourced availability of products. If a standard SFF solves unique design issues in a cost-effective manner, then it will gain market share.

FINSTEL: An abundance of embedded form factors, both standards-based and proprietary, has been introduced during the past five years. Each has its pros and cons and best-fit applications areas. The plethora of products, however, has made it more complex for designers to determine the optimal solution for a particular application. To that extent, yes, the market has become overfragmented and somewhat confusing to designers. It is important for embedded applications to have product longevity and design flexibility that enables generation-to-generation migration. Having too many form factors not only results in fragmentation of the market, but also potentially leads to the selection of a limiting solution and high total cost of ownership for the end design.

Even with the wide selection of standards and form factors, proven embedded computing suppliers that can help OEMs get to market faster with reduced risk and lower costs will still be able to gain market share.

LEE: While it may appear that fragmentation has occurred, each SFF standard addresses specific requirements that are market-driven and can greatly benefit a customer’s application. For example, the highly ruggedized “pin and socket” stackable interface of PC/104 is ideal for applications that require a reliable I/O interface in an environment with a lot of vibration. Alternatively, Computers-On-Modules (COMs) enable a customer with a proprietary SBC design and customized I/O requirements to create a solution that is both scalable and future-proof, should the CPU platform reach obsolescence. Additionally, new, highly integrated 32 nm CPU technologies allow a much smaller SBC footprint, thus creating smaller form factors such as Qseven and Micro COM Express. This benefits customers who have tight mechanical constraints and low power requirements. Because all of these form factors are based on long embedded product life cycles, a legacy effect happens in which the number of standards compound with the advent of newer standards being introduced.

Although an abundance of form factors might seem to limit market share for a given standard, the aggregate opportunity is greatly increased as newer standards appeal to broader markets and maximize a customer’s ROI with legacy solutions. It is why we believe the SFF market is still growing. The variety of SFFs is therefore a result showing the vendors’ efforts to try to grab as much of the market share as possible from this ever-growing market. The combination of new technologies, product features, and price point enables it to meet the different requirements of various applications, which cannot be fulfilled by previous-generation technologies. These new technologies include high computing performance, greater integrated video performance, low power consumption, wider temperature support, and several other System-on-Chip (SoC) features.

LIPPERT: I don’t think the SFF market is overly fragmented. As technology evolves, the need for form factors to leverage that technology will increase. New form factors such as CoreExpress (Figure 2) allow customers to select the one that is right for them instead of having to fit the requirements of a form factor that is not ideal for their applications. This allows companies to enter market segments that previously would have had high barriers to entry because the customer would have had to develop a custom board, helping fuel the growth of the entire embedded market. This does make it difficult for any one form factor to gain market share, but does not preclude any given company from growth if they build products that utilize a variety of complementing form factors.

22
Figure 2: New form factors such as CoreExpress, which was originally developed as a proprietary standard by LiPPERT Embedded Computers and is now an SFF-SIG specification, allow customers to pick and choose the modules that fit their requirements.
(Click graphic to zoom by 1.9x)

MILLER: The perceived fragmentation of the market can be rightfully viewed as an illustration of diversity in our world and a statement that one size doesn’t fit all. Although the market appears fragmented at first glance, the reality is all standards don’t have equal market share. The fact that customers continue to select SBCs with PC/104 (ISA bus) expansion reveals the dichotomy between the evolution of technology in the lab and the reality in the marketplace. It further indicates the continuing strength of the PC/104 ecosystem and the reluctance of customers to reengineer and requalify proven systems. As the saying goes, “if it ain’t broke, don’t fix it.” Customers are hesitant to migrate ISA to PCI or beyond without any meaningful payback. Our customers are asking us for new SBCs that work in their existing PC/104-based applications because that simplifies not only their technology refresh effort, but CE, FCC, and other recertifications as well. Diamond is not alone in preserving the ISA bus for 7-10 more years, as evidenced by dozens of new SBCs announced in 2009 and 2010 with PC/104 expansion capability.

Of course, there is a glut of board sizes, as so many vendors seek to go it alone and capture market share by defining what they see as the ideal form factor and then attempting to exploit their position by shutting out competitors. Vendors who seek to create proprietary standards are tightening the noose around their own necks, however, because customers are always looking to reduce risk, and there is nothing more risky than a proprietary standard. Our philosophy at Diamond is to work within the confines of open standards because they bring so many benefits to both vendors and customers as a result of their built-in efficiency and interoperability.

TOMASSON: Yes, the market is fragmented, with various form factors for SBCs, embedded motherboards, mezzanines, COMs, PC/104 variants, and all classes of embedded computing platforms. The market is also fragmented by processor options, Operating Systems (OSs), vertical markets, customization levels, customer types,  geography, and so on. The embedded computing market is incredibly complex, varied, and yes, fragmented.

Does fragmentation in a market make it difficult for individual companies to gain significant market share? Not necessarily. Fragmentation requires companies to select one of two strategies: either focus on a niche or provide a broad portfolio of products across the market. Whichever strategy a company chooses, it must have differentiators that make it stand out in the market. That’s how any company will gain market share.

When a company focuses on standards and compliance to the many form factors in the market, it hinders brand distinction. When a company focuses on more than standard form factors, such as simplifying development and deployment, supporting customers through the entire product life cycle, and solving specific customer pain points, a surplus of standards and sizes has no impact on the strength of that brand or that company’s relationships with its customers.

SFF: Why are there so many SFF standards? In your opinion, does the embedded industry need all these different sizes of modules, or does it hinder competition in terms of brand distinction?

MILLER: To a great extent, the proliferation of SFFs is the result of a healthy, pluralistic market founded on technology that is easily malleable to customer needs. Some proliferation of standards simply comes from the desire to gain market advantage by creating something incrementally better than what’s already available. Such activity often leads to the emergence of valuable ideas that can better serve the broader market. This Darwinian effect is to be expected and celebrated, because in the end, the customer wins.

However, many customers are simply buying a particular set of features (such as processor type or OS compatibility), regardless of the exact form factor. In this case, the customer’s priorities are exactly the opposite of a form factor-based approach. A form factor is basically three things: physical outline, mounting holes, and expansion connector. Nowhere does it mention processor architecture or any other technology other than what’s on the expansion connector. For customers who can obtain all their needs with a single board, a standard form factor is not important. Here, vendors succeed by providing the correct set of features rather than the correct board shape.

Moore’s Law (the natural evolution of technology in a highly competitive environment) can be cited as a contributor as well. As technology evolves, new formats spring up to take advantage of new value propositions involving size, speed, power consumption, and integration. For example, Diamond’s newest data acquisition circuit has 8x the sample rate in one-third the board space as our previous top-performing product, and at a significantly lower price point. This was made possible by continuing advances in semiconductor and packaging technology.

This type of game-changing evolution leads to new form factors such as FeaturePak (Figure 3), which was invented by Diamond specifically to take advantage of these evolutionary dynamics. The embedded industry needs new standards such as FeaturePak to pass on valuable size, weight, and power savings to device manufacturers in all market segments.

23
Figure 3: The natural evolution of technology encourages the development of new form factors such as FeaturePak, the base specification for Diamond Systems’ FeaturePak-DAQ1616 module.
(Click graphic to zoom by 1.8x)

TOMASSON: The variety and quantity of SFF standards developed over decades, with multiple forces at play. In embedded computing, the device built around the embedded computer has a long shelf life. OEMs understandably want to extend those life cycles as long as possible and tend to choose the same form factor as technology advances push components to the end of their life and new embedded platforms arise.

In the past, niche providers optimized and invested heavily in their proprietary form factors, then joined or started standards bodies to promote and evangelize them. Many of those form factors became de facto standards.

The real question is: Does an existing form factor ever fade away? Not with devices and OEM products that are deployed, unmanned, and operational for decades. Telecom service providers in the past used 40-year depreciation periods for equipment in the network. Try telling them it’s time to pull those operating units out and invest in an upgrade. Now that’s a hard sell. They prefer to maintain the units that come in occasionally, which leads to more and more standards in the field as technology advances to enable new capabilities that require form factor changes.

As an example, some companies chose to put the Intel Atom Z5xx processor into a standard COM Express form factor, with much hype and fanfare. However, the COM Express form factor did not expose all of processor’s technological advances. Other companies enabled all the processor’s new technology capabilities in a different form factor, still small and with very low power consumption, but without the limitations of the COM Express standard form factor. At that point, it is the customer’s decision to use all the new processor’s capabilities or purchase a standard form factor to avoid being tied to one supplier.

BURCKLE: The reason that there are so many SFF standards is due to the variety of customer-driven applications. It is impossible to dictate one standard that satisfies everyone; therefore, a multitude of options has appeared. These options in most cases are designed to serve a specific application category or product feature. Unfortunately, there are also products from the “not invented here” group of companies, which think that their engineers can always design a better “mousetrap.” One of the keys to success is to define a unique niche and then aggressively execute a product development program that capitalizes on its strengths to meet the requirements.

FINSTEL: We’re in a time of technology change. Each introduced standard has good intentions to improve application development and is jockeying for position as the best implementation. The market has a way of weeding out poor performers: the weaker solutions go away and winning implementations move forward. ETX as the de facto standard for semi-custom solutions is a clear example of this, as well as the proliferation of solutions choosing to follow COM Express (see Figure 4). The industry as a whole might not need so many standards, but the existence of so many does give niche applications an off-the-shelf solution that customers would otherwise have had to pay much more for as a full-custom solution. The abundance of offerings has no impact on brand distinction. That is determined by quality, availability, fit – the things that set one solution apart from the others.

24
Figure 4: The proliferation of ETX and COM solutions like the Kontron ETXexpress-AI demonstrates how winning implementations thrive in the embedded market while weaker solutions disappear over time.
(Click graphic to zoom)

LEE: The abundance of SFF standards has arisen out of market demand, technologies available at the time, and the legacy effect of older standards remaining active due to long embedded product life cycles and significant customer investment in those original standards. Since these form factors have been created specifically to address certain application needs, we feel that these standards are necessary to offer the best possible suite of solutions to a customer.

LIPPERT: The numbers of form factors in the market is determined by customer demand and evolving technologies. Form factors must change to meet the new interfaces being developed. As customers demand more features in smaller spaces, form factors evolve to meet the demand, and others that cannot support this go away. As more form factors enter the market, each must differentiate itself from the others already in the marketplace. As the embedded market grows, segments of the market will have differing demands that will favor one form factor instead of another. This will lead to more form factors evolving.

SFF: What are the advantages of working in such a crowded market with dozens of options for customers to choose from? What are the disadvantages?

TOMASSON: One obvious advantage for the customer is the ability to have multiple sources for a standard form factor embedded computer. This multivendor market gives the customer strong control in the purchasing process. For the manufacturer, one advantage is simplicity of design – following an existing standard can be easier than creating a new design.

Disadvantages for the manufacturer include difficulties in building economies of scale by producing many different form factors and managing a portfolio of products that complies with different standards. Another danger for the manufacturer is being forced to compete on price in a multivendor market for a standard product, leading to a commodity pricing model. Competing on price makes it very difficult for a manufacturer to highlight any capabilities other than operational scale.

LEE: The main advantage is to have many tools available to provide the best possible solution for a customer depending on the application requirements. Additionally, when a standard is widely accepted by many embedded computing manufacturers, customers feel safe that they are not risking their R&D investment on a proprietary design with limited availability and potential obsolescence issues.

Marketing efforts, while supported by different suppliers, are in fact promoting the SFF standards (versus other form factors) and the extensive and expanding applications these SFF solutions could benefit. If you consider all SFF players to be competing in one industry, there is the potential for greater bargaining power when facing upper-stream suppliers.

The downside is to have your product get lost in a sea of “me too” products with little or no differentiation. Customers are confused with all the different standards and even more so when including proprietary standards in the picture. In the past year, with more new standards being added into the market, we have spent a significant amount of time explaining why a claimed “compatible” proprietary standard is not a “compliant” one and does not work the same way as a standard one would.

Another disadvantage is we are somewhat under the tension of a potential price war, and this would be pushed to a greater extent when commercial- and industry-grade products are being compared side by side.

BURCKLE: A crowded market means more options. So whenever there is competition, there will be innovation, which is good for the customer. Innovation drives better features and lower prices as vendors are listening to what the market needs. For example, the new ultra-mobile CPUs and chipsets using the Intel Atom or VIA Nano families required a new interconnect technology for stacking modules. That is why the SFF-SIG defined Stackable Unified Module Interconnect Technology (SUMIT) to support these technologies for embedded systems rather than using the older server-class silicon. SUMIT is a unique blend that supports high-speed serial buses such as PCI Express and USB 2.0 while maintaining a bridge to legacy I/O technologies.

The disadvantages stem from the potential confusion and delays by designers as they try to sort through the options during the selection process. There are also business issues that need to be considered concerning the long-term financial and technical viability of companies, especially if they only offer a nonstandard solution from a single source.

FINSTEL: The key advantage is wide selection of options and solutions for designers. That advantage, however, is a double-edged sword. With such a diverse selection, it is much harder for a designer to do an apples-to-apples comparison.   The lack of compatible options makes second sourcing much more difficult, which impacts the life of the application and later, the ease of the migration path for next-generation solutions.

LIPPERT: The advantages of the vast array of form factors are the market’s ability to choose the form factors that best suit customer requirements and the ability to build a product that best fits the technology needs of the market you want to target. Companies can choose to develop form factors that best work with your targeted market segments. A disadvantage of a large number of form factors in the market is selecting form factors that do not align with potential customer requirements.

MILLER: Diamond benefits from a crowded market that rewards innovation. Paying attention to customer requirements and delivering better solutions has allowed us to grow. The fact that the supplier base is crowded only makes us work harder and more creatively, as witnessed by our product announcements from the past 12 months.

The disadvantage of a crowded market is that suppliers who cannot figure out how to add value, differentiate, or meet customer needs will suffer, especially in this weak economy.

SFF: How is your company differentiating its embedded modules and/or systems to create best-of-breed products that stand out in the market? Describe your design strategies.

LIPPERT: LiPPERT Embedded Computers has chosen to focus on areas of the market that have advanced requirements like extended temperature ranges and shock and vibration resistance. Our goal is to design every product to ensure it can handle the harshest environments. We differentiate our products by focusing on quality and customer service.

MILLER: Well, I will just say that Diamond seeks to devise the ideal solutions that balance technology, flexibility, and cost, while keeping our eye on the market segments that can bring the best chances for profitability. We don’t chase the highest-performance or the lowest-cost sectors at the ends of the risk bell curve. Customers appreciate simplicity, so our recent product introductions incorporate many innovations that provide significant benefit to customers by eliminating many of the headaches inherent in prior techniques. In addition, our 2-in-1 concept, which combines CPU and I/O onto a single board to reduce system size and cost, is well-known in the industry and has been well received by our customers.

BURCKLE: WinSystems targets the industrial, communications, transportation, medical, and military/COTS markets. As such, we design our products to withstand the environments in which they will be used and provide customers award-winning, no-cost technical support during their design phase and into production. We offer direct, live, engineer-to-engineer phone support as we partner with customers to make their designs successful. We also support only the x86-class processors. This allows our customers to work with Linux, Windows XPe, and other compatible Real-Time Operating Systems (RTOSs), drivers, and development tools in which they have experience. The result is a faster product introduction because of the wealth of software options available for them to quickly design, develop, and test their application programs with confidence in a short amount of time.

LEE: Portwell’s strategy is to differentiate at both the product and service levels. For our designs, we look beyond simply breaking out common I/O from an I/O hub. We try to understand additional feature sets that would appeal to a customer’s application, including harnessing technologies such as Intel’s Active Management Technology for remote system management or Intel’s Virtualization Technology for multiple virtualized guest OSs. And as a longtime member of the Intel Embedded partner program, we align our new products with Intel’s embedded roadmap, and our products come with seven-year life-cycle support. The strategic relationship with Intel brings Portwell early access opportunities, which bring shorter time-to-market advantages to our customers. We design, develop, and manufacture all our products in a state-of-the-art facility with high-tech equipment and strict quality management. Furthermore, we implement our vertical market know-how/experience and customers’ feedback in previous and existing applications and future product design and development.

For the services model, we differentiate by having a very flexible “high mix, low volume” customization policy allowing customers to outsource additional services with very competitive terms. We also offer extended engineering services (such as custom design, integration, and production) and dedicated logistics support (such as local hub and drop shipment) in addition to standard, off-the-shelf products.

TOMASSON: Eurotech differentiates our embedded computers and systems by focusing on our core strengths internally and by simplifying the entire development and deployment process for our customers.

Our embedded computers make the most of the latest technology from our processor partners and draw out all the capabilities they offer. When our customers are paying for the latest technology, it just does not make sense for us to restrict the functions available to them because an existing standard did not anticipate the capabilities. In addition, we take pride in pushing the power consumption down, farther than some might believe possible, and proving our development success in that area with unbiased testing.

Software is a third focus for Eurotech, and we have learned after years of satisfying customers that the out-of-the-box experience is vital to the success of the OEM developer. Our embedded computers come loaded with the OS so OEM developers can plug it in and start the application development quickly. With our latest Everyware Software Framework middleware (Figure 5), application development can be even faster due to its hooks for application developers.

25
Figure 5: To stand out in the crowded SFF market, Eurotech focuses on simplifying the development and deployment process through products such as Everyware Software Framework middleware.

Finally, our external focus on simplifying the entire product life cycle for our customers differentiates Eurotech in the market. From early development with ESF and North America-based support through end-of-life support and technology migration, Eurotech delivers the embedded computing platform that lets our OEM customers create and deploy their products quickly and efficiently.

FINSTEL: Kontron has a history as a demonstrated leader in embedded computing technology, and we have successfully developed solutions that have become industry standards. Our method is straightforward: it starts with understanding the designer’s immediate needs and requirements for the future. When it comes to standards developed by other organizations, Kontron takes the same customer-centric approach when evaluating each standard’s long-term viability to the market to determine which standards we will adopt and support.

While Kontron is global in strength and scope, we give customers local support. We have a keen focus on developing products that give our customers a clear competitive advantage in terms of processor, performance, power, reliability, and integrated features. Kontron has strong alliances with the major processor and software manufacturers, so we can consistently offer the most cutting-edge products ahead of other suppliers. Ultimately, the benefits for our customers come down to helping them make their time-to-market, development cost, and product longevity goals with the best product possible.

SFF: What is the value of developing standard-size modules as opposed to proprietary modules, or vice versa? How does open source fit into the SFF landscape?

BURCKLE: The value of any standard is that a product should and will be built to a known specification. Standards are defined to ensure conformity, compatibility, and interoperability. They become the keystone of an ecosystem of products developed by independent manufacturers plus specialty I/O boards by the customers. Over time, more and more products become available as the standard proves its viability. This breeds a larger market that fosters competition, which in turn spawns lower prices and better products with more features. For projects that can use a standard module, it is very cost-effective and reduces time to market since a project engineer is already familiar with its design attributes. Multiple supply sources are also important for long-term availability if one vendor is no longer able to continue delivering a product. There is safety in numbers.

Proprietary modules are a good choice if an application needs a unique form, fit, or function that cannot be found with current off-the-shelf modules. For example, I have seen round boards that are designed to fit inside a 6-inch pipe for petrochemical monitoring and control applications. Also, as its name implies, a proprietary approach might be needed if a company has intellectual property or other secret information that it does not want to disclose to the market or competitors. However, these modules require a brand-new design effort and are typically limited to a single vendor.

Open source is critical so that any manufacturer or OEM has the license- and royalty-free rights to develop products to meet the application requirements. By encouraging adoption of industry standards, there will be a growing base of products in that form factor’s landscape.

FINSTEL: It all comes down to design flexibility – capabilities and availability. When designers choose a standard module, they get scalability in terms of performance for today’s application that can also be applied for future generations. That same scalability and longevity is not guaranteed with proprietary solutions.

Standards-based form factors can also limit risks and help simplify or facilitate the development process, which lowers the total cost of ownership. Open sourcing also ensures that the benefits achieved with standards are maintained.

LEE: Since the beginning of the COTS revolution, the embedded computing market has greatly benefited from a standards-based approach. Customers have been able to greatly maximize ROI by creating stable, repeatable platforms in which many components can be reused while migrating to different CPU/chipset combinations. For example, a customer who designs an enclosure around a Mini-ITX motherboard (Figure 6) can typically reuse the same chassis design with another motherboard, as the mechanical layout will be the same. Additionally, a customer that develops an ETX carrier board can reuse the same carrier board design when upgrading to a different ETX-based module. All in all, the value presented is shortening the customer’s new product introduction process, minimizing the risk in a new design, speeding up time to market, being cost-effective, and leveraging marketing efforts so that more resources can be spent on promoting the product itself and less resources on the standard.

26
Figure 6: A standards-based approach allows customers to design an enclosure around a Mini-ITX motherboard like the Portwell WADE-8067 and reuse the same chassis design with another motherboard.
(Click graphic to zoom)

LIPPERT: The value of developing a standard module is you can leverage development that has already been done. You don’t have to determine if a given set of signals will be adversely effected by the connector chosen because this has already been validated for you.

MILLER: The primary value of standards is not in their size, but in their interfaces. Because the PC/104 connector is present on so many form factors, PC/104 module vendors have the opportunity to sell to a wide range of customers. This attracts more I/O module vendors to the market, which in turn makes the standard more desirable to CPU vendors, leading to more sockets, and the cycle continues. The value to the customer is a large selection of interoperable products from multiple vendors, leading to greater choice and lower risk. Proprietary modules have their place when there is a set of features that cannot be implemented any other way or when a technology is new and has not yet achieved wide adoption. In these cases, limited choice is acceptable to the customer because the priority is on the feature set.

Open source has a similar self-reinforcing dynamic in the standard module market. It creates a magnet that attracts developers, who in turn attract more customers. In addition, open source makes it easier for many companies to develop new hardware because the software part of the solution is already available and does not need to be reinvented.

TOMASSON: Eurotech has a balanced model of developing to standard form factors and designing the precise solution for certain customer requirements, so we see the value of developing both for the fragmented embedded market. The value of developing to the standards is to meet market demand in that area for those customers who are willing to accept more limited capabilities and options for their products. However, the most rewarding test of Eurotech’s engineering talent is to create the most efficient product in the smallest form factor with the lowest power draw and highest performance. Those embedded computers enable the OEM’s product to stand out in the market.

Open source is a vital aspect of the SFF landscape, and Eurotech is a strong proponent of delivering open source systems as part of the overall product portfolio. Our latest embedded computers and systems offer open source options in addition to other OSs, due in large part to market demand. Our ESF is based on open source fundamentals and brings additional value to the OEM market.

SFF: Do you think the number of SFFs will continue to proliferate, further fragmenting the market? Should it?

FINSTEL: No, I believe the explosion of SFF development will slow down. We will see that the strongest standard solutions become widely adopted by the mainstream market and most likely extended. Some of the weaker SFF solutions will go away completely, while a few might find a place in serving niche markets.

LIPPERT: I think the marketplace will continue to evolve, with more form factors being developed and others becoming obsolete. I don’t think this is a weakness of the embedded market. The market will listen to its customers. If customers adopt a large number of form factors, this is a statement of the diversity of customer applications and requirements more than a statement on the status of the industry. The market will continue to evolve to meet customer demand.

TOMASSON: Our perspective is that the number of SFFs on the market doesn’t matter. Hardware is the foundation, the starting point for the embedded market. With the latest connectivity and software capabilities, the hardware form factor is less important and will not be the defining factor for the success of an embedded manufacturer, especially for new device designs.

LEE: As CPU manufacturing processes continue to shrink beyond 32 nm and additional I/O resources will become available in these newer CPU/chipset technologies, the number of SFF specifications will continue to grow. While it may be construed as market fragmentation, this is necessary to harness newer technologies and provide more flexible solutions demanded by the market. Additionally, as the market completely adjusts to newer interfaces such as PCI Express from 32-bit PCI, you will start to see a consolidation of some of the older standards, thus reducing the amount of fragmentation in the market. So in answer to the question if it should be further fragmenting, we believe the market itself will determine what happens.

MILLER: Today’s highly integrated technology combined with sophisticated development tools keeps the barrier to market entry relatively low. This attracts new entrants into the market, which results in the continued proliferation of standards as these new players introduce their own ideas. This is a healthy situation because it continues to bring innovation, value, and possibility to customers and society. Standards organizations bring valuable sanity to the situation by providing a consistent playing field for vendors to work together and make it easier for customers to utilize all these available technologies.

BURCKLE: Newly created SFFs will continue to evolve and proliferate due to the changing nature of the processor/chipset interfaces and opportunities from the applications that they serve. One recent example is Pico-ITXe and companion Pico-I/O expansion modules introduced by the SFF-SIG (see Figure 7). This smaller form factor will enable a host of new space-conscious mobile OEM systems for new and growing computer monitoring and control markets. A Pico-ITXe system footprint is only 72 mm x 100 mm, with an even smaller 60 mm x 72 mm for the stackable Pico-I/O module. Enabled by SUMIT technology, a Pico-ITXe SBC can also support a wide variety of other external interfaces with a variety of I/O bus speeds.

27
Figure 7: Changes in processor/chipset interfaces will continue to spur the creation of new SFF standards such as Pico-ITXe and Pico-I/O, which are available in the EPIA-710 and PCO-UIO48-G offerings from VIA Technologies and WinSystems, respectively.
(Click graphic to zoom by 1.7x)

Additional SFFs will continue to be introduced as long as there is a legitimate need for them and emerging technologies to drive them. The difficult question to answer is: Will a new SFF fragment the market, or will it define something new that could grow the entire SFF arena in a significant manner?

Robert A. Burckle is VP of WinSystems. He has more than 30 years of experience in embedded computing and earned both Bachelor’s and Master’s degrees in Electrical Engineering from the University of Louisville plus an MBA from the University of North Texas. He can be reached at bburckle@winsystems.com.

Dirk Finstel is CTO of Kontron, where has led advancements in ETX and COM Express. He has worked in the industry for more than 18 years and holds a BS in Computer Engineering and Science from the University of North Carolina at Charlotte. He can be reached at dirk.finstel@kontron.com.

Allen Lee is CEO of American Portwell Technology. Allen has more than 15 years of experience in the industrial and embedded computing industries and holds a Master’s degree in Engineering Management from the University of Pretoria. He can be reached at info@portwell.com.

Peter Lippert is the founder and general manager of LiPPERT Embedded Computers. Under his leadership, the company showed continuous growth and today has more than 40 technical employees in Germany and the United States. Peter completed his MSEE in Telecommunications in 1979. He can be reached at plippert@lippertembedded.com.

Jonathan Miller is founder and president of Diamond Systems Corporation. After working in the PC-based data acquisition market, Jonathan founded Diamond Systems in 1989. As a charter member of the PC/104 Consortium, Diamond has introduced several new form factor innovations under Jonathan’s leadership. He can be reached at sales@diamondsystems.com.

Hilary Tomasson is VP of marketing for Eurotech North America, where she drives product management and marketing communications initiatives. She formerly led BroadSoft’s corporate marketing activities as the company grew to be one of Deloitte’s Top 100 Fastest-Growing Companies in North America in 2006. She can be reached at hilary.tomasson@eurotech.com.