Navy Acoustics-Rapid COTS Insertion program heads list of successes at COTScon West 99
SAN DIEGO Many military programs are employing commercial off-the-shelf (COTS) technology, according to participants in the COTScon West`99 conference last month in San Diego. Yet the one that got the most attention was the Acoustics-Rapid COTS Insertion (A-RCI) program to upgrade the U.S. Navy`s submarine fleet.
By John Rhea
SAN DIEGO — Many military programs are employing commercial off-the-shelf (COTS) technology, according to participants in the COTScon West`99 conference last month in San Diego. Yet the one that got the most attention was the Acoustics-Rapid COTS Insertion (A-RCI) program to upgrade the U.S. Navy`s submarine fleet.
Commenting on the A-RCI were representatives from the customer, Capt. Thomas O`Connor, A-RCI program manager at the Naval Sea Systems Command, Arlington, Va., and the prime contractor, Kenneth Moore, senior federal systems architect for advanced systems applications at Lockheed Martin Undersea Systems in Manassas, Va.
O`Connor and Moore detailed how COTS was saving money, reducing development time, and improving reliability through the use of open-system architecture and a "cocooning" or "armadillo" approach to enable the commercial-grade components to survive in a submarine environment.
The A-RCI program launched with the future Virginia-class attack submarines, O`Connor says. Within the next year and a half, he expects it to spread to 75 percent of the Navy`s submarines, including the Los Angeles and Seawolf classes.
The A-RCI resulted as Navy leaders faced two serious problems, he recounted: they were not realizing the full theoretical gain of their submarine sensors, and the existing 1970s technology was too costly to upgrade.
COTS was the solution, O`Connor noted, and by this he meant "state of the practice," not state of the art. For example, he noted that 400 MHz microprocessors today are adequate for the submarine missions; there was not need to try out new 800 MHz practices in operational scenarios.
"The budgets helped us to say no to modifying COTS," he recalled. Instead, by using the predefined open-system architecture, the Navy was able to get portable software from other programs and use it as "middleware" to upgrade onboard processing with standard VME boards, processors, and displays.
Another contributor to this success story was close coordination with the end users in what O`Connor called the "build-test-build" approach. The program proceeded in phases, beginning with towed array processing, then hull-based processing, and then on to upgrading the high-frequency components. All through the phases, he noted, "The fleet guys are part of this process."
From the producer viewpoint, Lockheed Martin`s Moore reported that use of COTS had yielded nearly a 3-to-1 improvement in mean time between failure (and 8-to-1 in mean time between critical failure), cut development cycle time by 2-to-1 (from the traditional three to four years to a year and a half to two years), and even reduced the power requirements on board the submarines by 4-to-1.
The net result, according to Moore, was a 4-to-1 reduction in total ownership costs and the ability to support the upgraded system for 15-to-20 years through technology insertion and technology refresh based on the standard form, fit, and function approach.
This is not a 100 percent COTS insertion, he added, but with proper partitioning he expected to achieve 75-to-80 percent COTS. A critical factor is the cocooning, which isolates the COTS parts from the shock-and-vibration that submarines endure, such as 50-G pressures from depth charges.
A side benefit of the COTS effort on the submarines is contracting out more of the work to leverage new technologies and keep Lockheed Martin`s own employees in the mainstream of technology.
In cooperation with the Virginia Institute of Technology in Blacksburg, Va., Lockheed Martin has developed a COTS Assessment and Selection Tool, or CAST, to evaluate vendors. This has also led to what Moore estimated as a 40 percent reduction in sparing by using the COTS hardware across different platforms.
The heart of the submarine system architecture consists of the Fibre Channel and asynchronous transfer mode protocols, he noted. As a result, his company has become a system integrator anticipating new technologies rather than trying to squeeze out the last bit of performance from existing technologies. "The primes used to sell iron, but now the money is in the cards and software and integration," Moore concluded.
Based on this example and his own studies, Joe Chapman, the former Texas Instruments executive, listed five points that potential COTS users should take into consideration:
(1) Continue to coordinate efforts through integrated product teams.
(2) Work with the system prime contractors and the component suppliers as early in the program design and definition phase as possible.
(3) Recognize that the life cycles in military systems vary considerably from those in the commercial, industrial, and consumer markets.
(4) Participate in such coordinating efforts as the Avionics Working Group, which includes such airframe manufacturers as Boeing Co., Seattle.
(5) Participate along with customers and vendors in DOD strategy and technology meetings.
Chapman, who surveys COTS standards for the Defense Standardization Program Office with his own firm, Chapman Consulting in Midland, Texas, also reminded the attendees that the military market does not have the clout it once had among electronics component suppliers. The companies, he says, will have to adjust to this situation.
One approach that enlivened past COTScon conferences — upscreening commercial parts by testing them until some were found to meet military specifications — is still an unresolved issue because the Defense Department has still not established a uniform policy, he noted.
Studies continue at several DOD-sponsored working groups, Chapman added, and the Naval Surface Warfare Center in Crane, Ind., has put out its own requirements and application manual, which he said was a useful reference. However, industry is still waiting for a memorandum that will give guidance on how to use upscreening based on all the studies to date.
Without belaboring the point, Chapman contended that systems designers should still consider upscreening only as a last resort. "There`s no question it`s been going on for 10 years," he said
Chapman noted that damage from handling parts during the upscreening process has been established as the largest contributor to defects and that an upscreened lot of parts is still only what he called "a snapshot in time" that is unlikely to be repeated in tests of subsequent lots.
Ray Alderman, executive director of the VMEbus International Trade Association (VITA) in Scottsdale, Ariz., offered his own appraisal of COTS in what he called the "mean time to obsolescence," or MTTO.
His formula for determining the MTTO of a given board-level system is to divide the observed life cycle of a chip technology by the square root of the number of unique chips on the board. For example, if there are three chip types on a board and the lifecycle is 15 months, the MTTO is 15 divided by 1.732 (the square root of 3), or 8.66 months.
The decision on whether to use COTS is an operations research problem, Alderman related, that involves such concerns as cooling, power distribution, and RFI/EMI considerations.
As a proponent for VME boards, he expressed satisfaction in the growth of that technology in the past year, from $1.26 billion in sales in 1998 to $1.4 billion in 1999, with the military remaining the principal customer.
Alderman also offered some wry alternate definitions of COTS. From the end user`s point of view, he said, COTS means Can`t Operate To Specifications. Manufacturers using COTS parts call it Continually Obsoleting The Software, and vendors disenchanted with the minuscule military market refer to it as Concentrate On Telecommunications Sales.
Among the other successful users of COTS, Amos Deacon, chef executive officer of Phoenix International in Orange, Calif., reported that open systems architectures had permitted the continuous upgrading of the rugged mass storage systems for the U.S. Air Force`s Joint STARS, Compass Jammer, and Rivet Fire aircraft programs dating all the way back to 1981. Flat panel displays and VME racks were added as needed.
Commercially available fiber optics have been successfully integrated into a variety of military systems, according to Dennis Horwitz, vice president of the Mil/Aero Division of Rifcos Corp., Camarillo, Calif. These programs include shipboard RF antenna links and the Phalanx close-in weapon system for the U.S. Navy and the U.S. Air Force`s Airborne Warning and Control System upgrade with gigabit Fibre Channel local area networks, plus in-flight entertainment systems for commercial airlines.
The advantages of fiber optics are well known, Horwitz noted, including reduced weight, greater bandwidth, EMI/RFI isolation, and elimination of fire hazards. Yet he also urged potential users to design their systems to take maximum advantage of fiber optic capabilities rather than installing the new technology as a one-for-one replacement of existing copper cables.
Also at the first COTScon on the West Coast, Ted Glum, director of the Defense Microelectronics Activity (DMEA) in Sacramento, Calif., reported that the increased reliance on precision and other "smart" weapons across all the military services dictated a common approach rather than each service tackling the problem on its own.
Among the DOD-wide approaches DMEA is evaluating is to move to lower power levels — from traditional 5 volt systems down to 3.3, 2.5, and eventually to 1.8 volts. Many of the COTS parts the defense industry needs are available from reputable aftermarket suppliers, Glum added, but he warned against buying from what he called "Internet brokers" lacking a known track record.
DMEA, although originally sponsored by the Air Force in 1981, is attempting to coordinate efforts among all the services and their prime contractors and represents what Glum called the "only horizontal organization" of its type within DOD.
In addition to an in-house flexible foundry to make obsolete parts ranging from mixed-signal components to radiation-hardened microelectronics, and multichip modules and hybrid circuits, DMEA is also linking users with suppliers from the aftermarket and semiconductor industries. The lead time for obtaining parts has shrunk from more than six weeks to 23 working days, Glum claimed.