SMALL FORM FACTOR FORUM
The Emperor's New Clothes
COLIN MCCRACKEN & PAUL ROSENFELD
Page 1 of 1
News flash: The latest SFF standard has been approved, and products are rolling off the line soon. Order your sample or dev kit today!
Not so fast. How do you know this new “thing” will stand the test of time? What if the spec was drafted by a lone technical guru, rubber-stamped by only a handful of others, and then launched under the pretense of industry-wide support from a respected trade group? Would that perspective lessen your design-in urge? Do we compliment the Emperor on his or her new clothes, or form an opinion based on our own thorough technical and market evaluation?
To analyze this consider how Corporations win or lose in competitive markets based upon the relevance of their offerings. The market decides. What would happen if your marketing folks defined products without regard for what can actually be built? At the other end of the spectrum, what would happen if your engineers designed products in a vacuum without market input? The days of “build-it-and-they-will-come” are over. Successful new ideas come from an engineering / marketing partnership where available technologies are applied to real customer needs.
Sounds simple. Yet the small form factor community is discovering déjà vu all over again. Processor and chipset vendors have appealing new products, and updated standards are needed to take advantage of the new bells and whistles. Whether board-level or SSD-level, a number of trade groups appear to have created new standards incorporating new features that embedded system OEMs don’t need or want simply because a processor or chipset offers such a feature. Can you say, “type 2” or “type 6”? What is going wrong?
The value of a standard derives from the ability of customers to apply a variety of compliant products to solve design, manufacturing and lifecycle management challenges over time. Standards that do not meet the needs of a particular target market likely won’t stand the test of time. However, in this market it takes many years to determine success or failure of a standard, especially if the solution looks far forward into the future. Marketing professionals are always refining their view of target customers based upon perceived current requirements (the famous moving target so detested by engineers everywhere) and upcoming challenges to reduce size, weight, power, cost and so on. Trade groups could benefit from the same level of market analysis that system OEMs routinely use.
In creating these new standards or pinout variations, it is very tempting for engineers merely to look at the latest chipsets and map the buses and I/O to off-board connectors. But this tends to disregard the installed base. The entire set of signals can be massive overkill, even for mainstream applications. High-density connectors allow many more pins in the same space. It doesn’t mean we need to use them all. Smaller connectors with a well-chosen pinout provide the opportunity to shrink overall system size. Consumer flash modules offer greater data bandwidth at the cost of higher power consumption, but in some cases these modules are literally too hot to handle by embedded OEMs.
A standard specification is the product of the input, creation and review process of a trade group. It has to stand the test of the market, regardless of the marketing hype surrounding its introduction. Results can range from acceptance of a carefully tested simple migratory step to an un-validated misfire, and all shades of gray between. Diversity of thoughts and ideas is critical to the development process. The embedded community must not be afraid to engage in painful debates during the standards creation process about what features must stay and what can go. There just isn’t enough space, cost and power for kitchen-sink solutions.
Users of standards-based products are faced with many choices for next-generation designs, upgrades and retrofits. To make truly informed decisions, each OEM must research potential solutions using as many independent sources as practical. Naturally, suppliers will position their products in the most favorable light possible, so one must dig deeper. Don’t assume fitness for use given prior successes or reputation of a trade group. Check references and independent articles, evaluate standards against system-level requirements, and then choose wisely. Irrational exuberance doesn’t guarantee winning system designs. So don’t kiss up to the Emperor. Join the debate and be prepared to share your requirements rather than gush about the latest spec or pinout type. Get involved!
As usual, comments about this topic can be mailed to firstname.lastname@example.org.