Evolution to COTS: Yes and maybe
Walsh, Edward J
The Launcher Systems Branch of the Dahlgren Division of the Naval Surface Warfare Command (NSWC) in Dahlgren, Va., plans to test the capability this spring of an experimental shipboard launcher, called the concentric-canister launcher (CCL), designed for use in launching torpedoes from surface ships. To generate the power needed for the launch test, the CCL test team will use an airbag inflator taken from a 1995 Dodge Caravan -a device that, according to the program team, creates an intense burst of pressure-only for an instant, but long enough to pop the torpedo into the water.
Besides determining the potential usefulness of the CCL for torpedo launch, the test also reflects the wide range of applications of commercialoff-the-shelf (COTS) products to support, and in many cases replace, Navyunique systems. Current Department of Defense acquisition policy, described in DOD Instruction 5000.2, and reflected also in Navy policy, directs the services to use performance specifications for purchasing new systems and/or for major modifications and upgrades; when government (i.e., military) specifications are not practicable, a nongovernment, or commercial, specification may be used. Both approaches steer the services hard towards commercial products. DOD 5000.2 also stipulates that waivers are necessary in most cases for the procurement of military-unique products, even for those purchased for integration with tactical systems.
Higher Tech, Lower Cost
The Navy is pursuing numerous initiatives to insert COTS technologies in shipboard, undersea, airborne, and command, control, communications, computer, and intelligence (C4I) systems, not only to take advantage of the significant savings achievable through COTS, but also because most commercial systems, developed to respond to the fastchanging civilian markets for technology, are nearly always closer to the state of the art than Navy-unique systems. For example, Lockheed Martin Government Electronics Systems in Moorestown, N.J., longtime prime contractor and systems integrator for the Aegis weapon system, will soonpossibly as early as this month-select a commercially developed processor for the newest Aegis baseline that not only will not be a Navy-unique product, but also will come from outside the Navy’s “family” of commercial-based shipboard display processors, designated the UYQ-70 advanced display system or Aegis display system.
Lockheed Martin Tactical Defense was awarded the current UYQ-70 contract in January 1994. The company then was part of Unisys; it later was acquired by Loral, which subsequently was purchased by Lockheed Martin. The UYQ-70 award symbolized a decisive break with the practice of purchasing military-specification (MILSPEC) shipboard displays. An unusual feature of the UYQ-70 award was the inclusion of a Navy provision to recompete the contract three years later in order to maximize the insertion of new technology. A second Q-70 award is expected this spring; a team led by Litton Data Systems has been competing with Lockheed Martin for the new production contract.
Sequences and Sensors
The Navy’s shift to increased use of COTS technology for shipboard systems prompted an intense debate in the early 1990s among ship-systems program officials and engineers about the degree to which commercially developed computers, input/output devices, and interfaces could handle the missioncritical detect-track-engage sequence for shipboard air-defense. That role is managed aboard most Navy surface combatants by the Navy-unique UYK43 shipboard computer, first introduced in 1983. The UYK-43, which is fitted with 64 input/output ports for linkage to weapons, sensors, and peripheral devices, manages the Aegis commandand-decision and weapons-control systems and the Aegis SPY-1 radar, and is the current production variant of a series of MILSPEC shipboard computers introduced in the 1960s.
Aboard Aegis ships, the UYK-43 is configured specifically for the primary Aegis mission of wide-area battle-group defense against fast-moving antiship missiles, which demand real-time system performance based on the foolproof transfer, at subsecond speed, of sensor data to weapon systems. A key feature of the ’43, critically important for combat-systems management, is its ability to tolerate faults by reconfiguring the flow of data in the event of a system failure.
Unlike newer open-system computers that can run a wide range of software programs, UYK-43 is a “closed” system that runs only the customized Aegis tactical executive system (ATES) programs and other Aegis software, written in the obsolescent CMS-2 language. Nonetheless, no commercially developed computers, even today’s faster and much more powerful systems, have been able, since the Navy began purchasing the UYK-43 from Lockheed Martin Tactical Defense, to match the ability of the ’43 to manage the flow of target-track and weaponsengagement data. That capability actually is more important for combat systems than raw processing power and memory capacity. Rueben Pitts, director of surface combatants at NSWC/ Dahlgren, says that the UYK-43 “is like a storm drain, while commercial computers [with much greater memory storage space] are like swimming pools.”
Aegis Missions Expand
The introduction of new missions for the Aegis fleet-primarily the theater ballistic missile defense (TBMD) mission and management of the cooperative engagement capability (CEC) system that generates a consolidated track picture based on the data provided by a CEC network of air-defense sensorshas exhausted the capability of the UYK-43. The program executive office for Aegis/surface combatants plans soon to introduce-for Aegis baseline 6 phase 1, targeted initially for DDG 68 through DDG 78-an architecture of “adjunct” commercial processors that will support the ’43. When baseline 7 phase 1, planned for fleet introduction with DDG-91, becomes operational, the ’43 will be replaced by a network of distributed commercial processors.
The insertion of commercial technology for shipboard applications also is moving ahead for the Navy’s nonAegis fleets. The program executive office for theater air defense is consolidating the ship self-defense system (SSDS), developed by Hughes (now Raytheon) Naval and Maritime Systems for aircraft carriers and Wasp-class and Whidbey Island-class amphibious assault ships, with the Block 1 variant of the advanced combat direction system (ACDS).
The SSDS links sensors and weapons to a distributed commercial processor across a commercial fiber-optic local area network. The system processors are housed in LAN access units and provide control for individual weapons and sensors. The ACDS, which started development in the 1980s, is run by two UYK-43s. Raytheon has been studying approaches to porting, to the commercial processing architecture, the CMS-2 software that runs on the UYK-43s. The two systems also will be integrated with the CEC to produce a new combat system, designated AKCITA, for the nonAegis combatant fleet.
The rapid shift to COTS for use in shipboard combat systems has not ended the debate over the reliability of commercially developed products for mission-critical tasks. It has, though, resulted in greater attention being given to traditional approaches to systems engineering to achieve a prudent degree of integration of non-military components in critical Navy systems.
Learning From Failures
In the early 1990s the Defense Advanced Research Projects Agency initiated a project to evaluate several commercially available computers for military use; COTS computers thus might be selected, on the basis of processing speed and memory capacity, to carry out demanding combat-systems tasks. The Navy formed a cooperative agreement with DARPA for the project, designated the high-performance distributed (Hiper-D) program.
A Hiper-D laboratory was set up at NSWC Dahlgren. In Hiper-D demonstrations, NSWC managers noted that, when subjected to the stresses of shipboard time-critical combat-systems management tasks, the commercial computers invariably failed to demonstrate the same data-management capabilities possessed by the UYK-43.
The Navy-DARPA lab approach in recent years has stressed the evaluation of systems and technologies developed by industry and service laboratories in order to determine how, and if, commercial systems could be modified and integrated with others, including military systems, to meet fleet requirements.
Outside the combat-systems arena, the Navy is seeking to replace Navycustomized ship machinery with commercial technology. Several “Smart Ship” demonstrations, carried out aboard the Aegis guided-missile cruiser USS Yorktown between 1995 and 1997, evaluated the use of commercial systems for, among other shipboard tasks, bridge management, maintenance of ship machinery, and damage systems control. The program executive office for Aegis/surface combatants plans to award a contract this spring for replacement-with seven of the systems demonstrated aboard the Yorktown-of the engineering-control-system equipment now on board the Navy’s Baseline 1 Aegis cruisers.
COTS and Consequences
The introduction of COTS components to the Navy’s C4I and information-systems domain has been less controversial. Because C4I functions do not operate with the real-time speed required for shipboard air defense, the development of new Navy C4I systems has relied primarily on commercial technology. A decisive first step was taken in the early 1990s by Vice Adm. Jerry O. Tuttle, author of the Navy’s “Copernicus” vision, when he initiated a series of Navy purchases of COTSbased workstations (initially for use only in nontactical applications). The intent, as Tuttle and many others pointed out, was to exploit the geometric rate of growth in commercial processing power, which was then estimated to be doubling roughly every 18 months.
In January 1995 the Naval Telecommunications Command awarded a three-year contract to an industry team of Hewlett Packard, Harris Corp., and SAIC for a fourth-generation tactical workstation, designated TAC-4, that would run the Unix operating system. The award followed an intense competition that was drawn out by changes in Navy requirements.
Following the award, other bidders, including Hughes Information Systems and Digital Equipment Corp., said that the cost of going through a contract competition for a new workstation contract on the Navy’s accelerated timetable had become so high that they would have to think very seriously about competing again.
Under the contract, Navy, Marine Corps, and Coast Guard commands were able to purchase TAC-4 systems through the General Services Administration, and to specify precisely the system configurations they desired. TAC-4 and TAC-3 systems have since been installed aboard many Navy ships.
Windows to the Future
The continuing evolution of technology towards the Windows operating environment, with which many new Navy recruits already are familiar, reduced fleet interest in the sophisticated but complex Unix-based TAC-4. In January 1997 Adm. Archie R. Clemins, commander in chief of the Pacific Fleet, called for Navy computer users to shift-“except in very rare cases,” Clemins cautioned-from Unix to the Windows NT operating system developed by Microsoft.
The move away from Unix and to Windows NT prompted protests from some industry officials, several of whom questioned the wisdom of endorsing a particular commercial product in an area of technology in which new products are being introduced so rapidly. The Navy has not published an official directive specifying the use of Windows NT for particular uses. But all major U.S. vendors of workstations, including Hewlett Packard, systems integrator for the TAC-4, now offer both Unixbased systems for high-end applications, and for desktop applications, computers that run Windows NT.
With the expiration of the TAC-4 contract on 18 January 1998, the acquisition of COTS computing equipment has shifted to marketplace-based practices, based on the use of innovative purchasing tools called blanket purchasing agreements (BPAs), which are issued to selected companies by the Naval Information Systems Management Center (NISMC) and administered by the General Services Administration.
The BPAs enable Navy commands, as well as non-Navy government organizations, to purchase computer systems from companies solely on the basis of price and performance-without a formal contract vehicle. Because of the flexibility of the BPA approach, the Navy last year canceled plans for a TAC-5 contract competition.
NISMC issued BPAs last year-to Sun Microsystems Federal, Digital Equipment Corp. (now being acquired by Compaq), and Hewlett Packardfor purchase of a wide range of computing systems. Four more BPAs, called enterprise solution agreements and covering technical-support services, have been issued to SAIC, Computer Sciences, Lockheed Martin, and Litton PRC.
The Navy’s shift towards wider use of commercial systems has moved on a steady but not always straight path. Rear Adm. John Gauss, now commander of the Space and Naval Warfare Systems Command, noted recently that 37 different versions of Unix are available-and they are not interoperable. In addition, federal information processing standards (FIPS) for relational databases do not reflect the performance achievable by the software available from such companies as Sybase, Informix, or Oracle.
Gauss also points out, though, that the products of those companies are not always interoperable; software written to achieve high performance in an Oracle database will not work with Sybase without the use of translation software called middleware-“another opportunity for the government to spend money,” he says.
For certain commercial products, such as high-end software, Gauss says, the Navy faces the continuing challenge of determining, at the risk of purchasing products that require extensive and expensive modifications, to what extent they are authentically interoperable. The danger, he adds, is that the Navy may end up purchasing products that may comply with federal standards but nonetheless are “vendor-unique.”
The solution, he says, is for the Navy to compete at the lowest price for interoperable products and let the companies that win have the business.
Edward J. Walsh is the editor of Naval Systems Update. For subscriptions, write to P.O. Box 2697, Woodbridge, Va. 22193; telephone (703) 490-1151; e-mail: firstname.lastname@example.org.
Copyright Navy League of the United States Apr 1998
Provided by ProQuest Information and Learning Company. All rights Reserved