Small form factors: A new SIG in town

Story

January 13, 2008

First question: What's the most popular open standard board type used in military systems? Answer: the personal computer, both in desktop and laptop sizes. But if we narrow the question to "rugged, deployed systems," you'd primarily find 6U VME boards, followed by 3U CompactPCI, SEM-E, and then a bunch of other types too numerous to mention. For the past 15 years of the COTS era, VME's anything-but-small size has dominated rugged, deployed systems.

First question: What's the most popular open standard board type used in military systems? Answer: the personal computer, both in desktop and laptop sizes. But if we narrow the question to "rugged, deployed systems," you'd primarily find 6U VME boards, followed by 3U CompactPCI, SEM-E, and then a bunch of other types too numerous to mention. For the past 15 years of the COTS era, VME's anything-but-small size has dominated rugged, deployed systems.

But VME has limitations in power (too much), interconnect (parallel bus), physical size (approximately 6" x 9"), and cost, forcing mil designers to increasingly look at other options in all but the harshest systems. With an eye towards PC-based, low-cost, modular, industrial and military systems, the brand-new Small Form Factor Special Interest Group (SFF-SIG) will be ready to go public with their plans in a few months. We got a special sneak preview that I thought would interest our readers. If successful, the results of the SIG’s efforts may wind up in your next military system.

The companies behind SFF-SIG are still in stealth mode because many participate in other standards bodies, some of which may perceive the SFF-SIG to be competition. Of these, the PC/104 Consortium has probably the most to lose since 100 percent of the SFF-SIG companies do or have participated in the PC/104 Consortium in its 15-plus year history2,3. The companies kicking off the SFF-SIG include Octagon Systems, Samtec, Tri-M Systems, WinSystems, and VIA. According to the SIG’s president, Colin McCracken, there are three board members (VIA, Octagon, and WinSystems), seven members, and more in discussion.

SFF-SIG has three working groups ‚Äì SBCs, Stackables (like PC/104 or PMC mezzanines on VME), and Computer-On-Module (such as PICMG's COM where the CPU resides on the mezzanine card plugged onto an I/O-laden carrier board). This year, the SIG plans to work on all of these three areas with a special emphasis on a new stackable spec it won't reveal until April. In the meantime, the SFF-SIG is following in the PC/104 Consortium‚'s footsteps with its "Adopt a Spec" policy whereby existing SFF board types of value become standardized for multi-vendor adoption and market competition.

As memory serves, this is how EBX, PC/104-Plus (which added PCI to the ISA bus), and EPIC became PC/104 Consortium standards some years ago. I maintain a list of more than 60 different small form factors "some open, most proprietary" which should provide fertile ground for open standards adoption by the SFF-SIG. According to McCracken, some of those SFFs are already under discussion by the SFF-SIG. This is particularly exciting news for low-power and low-cost defense systems, as many of these SFFs were designed for semi-rugged consumer applications such as long-battery-life handsets orvdroppable portable mobile devices.

Key to achieving the SIG's mundane-sounding goals will be a combination of focused marketing efforts and a VME VITA-like rulebook that crisply brings topics up for a vote so that they don't become "over-engineered, academic corner cases,‚" says McCracken. We like the sound of that. One of VME's strengths has long been its VITA Standards Organization (VSO), which is aggressively run both technically and administratively. Additionally, the SFF-SIG recognizes that it's market success that matters over technical success. Hence, the charter focuses on rapid ecosystem growth with a divide and conquer plan to combine the strengths of worldwide leaders.

On the technical side, right now the group is focused on creating a slightly larger-than-PC/104 SFF that is I/O-centric. Part of the problem with many of those 60 SFFs I mentioned earlier is that they were designed around the CPU (AMD, ARM, Intel, VIA, pick one), and the CPU's bus also becomes the board's bus. If the CPU changes, then the standard is in trouble.

Instead, the SFF-SIG wants to make its first new standard I/O-centric, expressly looking at forward and backward migration between processors and even other SFFs. This means that the proposed new standard would accommodate many of today's (and tomorrow's) CPUs and peripherals, while paying attention to system aspects of tech refresh and spiral insertion strategies. For the military, this is translated to code portability and Pre-planned Product Improvements (P3I). This is a page right out of VME's rule book, and one that resonates very well in defense because 10-year upgrades are just business as usual.

But small form factors come and go, as do their sponsoring companies. The same might happen to the SFF-SIG. Will it be around next year or the year after? We don't know, but we hope so. Personally, I like what I'm hearing. I can't wait to tell our readers about their first SFF when it's unveiled on April 8, 2008.

Chris A. Ciufo
[email protected]

1. Full disclosure: I personally helped to catalyze the formation of the SFF-SIG, although I have no direct interest in it. OpenSystems Publishing, Military Embedded Systems’ publisher, is an unofficial sponsor of the SFF-SIG.
2. For more on the PC/104 Consortium, check out OpenSystems Publishing’s PC/104 and Small Form Factors magazine at www.smallformfactors.com.
3. By the way, VITA and PICMG memberships overlap by at least 40 percent.