Databus evolution: 1553 switching over to Fibre Channel
High-performance databus technology represents a spectrum of choices for military users. At one end is the proven MIL-STD 1553, which is just that: a military standard. At the other end is the emerging Fibre Channel, a commercial off-the-shelf (COTS) technology with extraordinary promise. In between are several candidates, most of them from the commercial world, such as Fiber Distributed Data Interface/Copper Distributed Data Interface (FDDI/CDDI), the IEEE standard 1394 (better known as Firewir
By John Rhea
High-performance databus technology represents a spectrum of choices for military users. At one end is the proven MIL-STD 1553, which is just that: a military standard. At the other end is the emerging Fibre Channel, a commercial off-the-shelf (COTS) technology with extraordinary promise. In between are several candidates, most of them from the commercial world, such as Fiber Distributed Data Interface/Copper Distributed Data Interface (FDDI/CDDI), the IEEE standard 1394 (better known as Firewire), and predictable enhancements to Ethernet. Some mixture of these technologies will constitute the databus business for years to come.
On one point everyone is agreed: 1553 is stuck at the 1-megabit data rate, and it will never be any faster. Attempts to implement a fiber optic variant known as 1773, which offered 100-megabit data rates, but in practice delivered about 50 megabits, never won wide acceptance. Somewhat paradoxically, in an industry where technology generations are frequently measured in months, the 25-year-old 1553 remains sovereign. Moreover, as a military standard, there are few realistic opportunities for 1553 to be designed into commercial systems to foster economies of scale and technological enhancements.
Fibre Channel, on the other hand, offers breathtaking vistas of multi-gigabit data rates and plummeting costs based on widespread commercial acceptance. The technology is open-ended and expands opportunities for military applications beyond even the Defense Department`s current futuristic plans, such as the Army After Next.
How much is enough?
The crux of the tradeoff, then, boils down to the question: is one megabit enough? The conventional wisdom is that if some bandwidth is good, more bandwidth is better. In effect, build it and they will come. But this represents the brute force approach to the problem, and maybe clever system architecture can wring more performance out of 1553 without taking the expensive and time-consuming step of introducing a new architecture.
This is difficult for the new clean-sheet-of-paper designs such as the U.S. Navy`s future Virginia-class attack submarine (formerly known as the New Attack Submarine), the U.S. Army`s Future Scout Cavalry System, and the multi-service, multinational Joint Strike Fighter (JSF). Yet the problem becomes even more intractable when it comes to retrofitting existing systems.
What seems the prudent approach is to build on today`s 1553 technology and slowly, carefully move toward the databuses of the future - to Fibre Channel or to one of the other technologies in the middle of the spectrum. That is the approach of ILC Data Device Corp. (DDC) in Bohemia, N.Y. "1553 has a long legacy behind it and a long future ahead of it," says Steve Friedman, DDC`s director of strategic products. He says 1553 should last at least another 20 years.
DDC is one of the stalwarts of 1553 and understandably has a vested interest in keeping that technology alive. The company continues to invest in enhancements and continues to introduce new products. New ranges of DDC 1553 applications include the radiation-hardened BU-61582 Space Advanced Communication Engine (called SP`ACE) terminal for satellite applications and the BU-61588 5 volt Miniature Advanced Communication Engine (Mini-ACE).
Curiously, Friedman notes, what was once the top-of-the-line databus is now migrating toward the bottom in terms of performance.
He cites the growing market for smart munitions as the driving force here. They will need databuses to integrate information to operate the wing tips and fins, and make course corrections based on Global Positioning System (GPS) data. Thus, while the new commercial technologies are set for inclusion in new complex systems, the old standard 1553 looks increasingly attractive for the small, cost-sensitive applications.
Projections among DDC analysts for their own markets reflect this sorting out of the databus business. DDC analysts have identified five segments (in ascending order of average selling price, or ASP) as: low-cost devices for munitions, roughly in the $150 range, 15 percent; mainstream databuses for media applications, 57 percent; bus controllers, 20 percent; remote terminals, 5 percent; and space components at the $1,300-to-$5,000 level, 5 percent. "As the ASPs go down, the units [numbers] increase," Friedman says.
Nonetheless, he sees DDC`s business evolving toward Fibre Channel from its 1553 base, and the company is scheduled to introduce its own Fibre Channel products late next year. "A unified network concept across military platforms is a reality," he says, to reduce costs and take advantage of the increasing processor power to connect databuses. "Fibre Channel lends itself to this very well," he says.
What Friedman calls the "big trophies" for Fibre Channel are the Virginia-class submarine and the JSF. He regards them as 10 years out, but this poses no problem. Fibre Channel, he explains, is sufficiently scaleable to progress up to 1, and then 2, 4, and 8 gigabits per second. He also envisions Fibre Channel applications in such existing systems as the U.S. Army Boeing AH-64 Longbow Apache attack helicopter, the Navy F/A-18E/F fighter-bomber, and the Air Force F-15 jet fighter and B-1B bomber upgrade, plus opportunities in the secret world of reconnaissance satellites.
Another 1553 stalwart is UTMC in Colorado Springs, Colo. Ron Hehr, strategic marketing manager, agrees that 1553 does the job for a majority of applications. "For legacy systems it`s difficult to make wholesale changes in the architecture," he says, noting that command and control systems in particular can exist happily with the 1-megabit limitation.
"None of the commercial specs are suited for real-time applications," Hehr maintains. "That drives the avionics people crazy."
For example, all three aircraft of the generation of the future Air Force Lockheed Martin F-22 Raptor jet fighter, which includes the Army Boeing-Sikorsky RAH-66 Comanche scout/attack helicopter and the canceled Navy A-12 attack bomber, were supposed to use fiber optic high-speed databuses. Congress canceled the A-12 after a series of daunting cost overruns. Even the JSF is likely to use a mixture of databus technologies, Hehr predicts.
However, Hehr also sees areas where neither 1553 nor Fibre Channel has the edge. The internal local area networks (LANs) of ships are well suited for Ethernet variants and FDDI, he says, and spacecraft in particular can benefit from Firewire since there are no legacy systems to dictate the choices. Hehr agrees that smart munitions represent the biggest opportunity for 1553 since, as he puts it, "there are more bombs and missiles than there are aircraft." Accordingly, UTMC officials are marketing their SUMMIT line of products involving multichip modules and transceivers, as well as memory optimized to drive down costs. Cost is the big constraint, Hehr maintains.
Another 1553 stalwart making the migration to Fibre Channel is Systran Corp. in Dayton, Ohio. Tom Bohman, products strategy manager, says his company is already producing units in both areas, including a new 16-by-16 switch for use in avionics laboratories that develop operational 1553 systems. Systran does not supply operational 1553 systems. The new device, known as the Multiplex Bus Switch, enables engineers to attach as many as 16 of their line replaceable units on one side and as many as 16 databuses on the other to dynamically control unit test, explains Joe Williams, the company`s marketing manager.
Systran officials also recently introduced PCI and PMC versions of the 1553 interface as well as a graphical user interface software package. "We at Systran actively support MIL-STD-1553 and believe it has a considerable future as a military and commercial avionics/vetronics databus," Williams says. "We believe that MIL-STD-1553 will continue to be a viable databus."
Nonetheless, Systran is also joining the Fibre Channel party with its FibreXpress network interface card operating at a 100-megabyte-per-second data rate and compatible with such architectures as PCI, PMC, VME64, and SBus.
"Fibre Channel is a robust technology used by the commercial world, and the government doesn`t have to pay for it," Bohman adds. He sees opportunities in the F/A-18 upgrade and the Virginia-Class submarine, particularly for demanding sonar functions, as well as for the Air Force`s Airborne Warning and Control System aircraft upgrade.
Putting the issue in perspective, at the left end of the spectrum 1553 is well understood and defined, yet as a military specification, it is not cheap, points out Doug Patterson, recently named marketing manager for military systems in the Fairfax, Va., office of SBS Embedded Computers. Patterson estimates that 1553 represents 99 percent of the databuses on deployed systems. Fibre Channel, on the other hand, "offers a gigabit right out of the chute," Patterson adds. "The large bandwidth is the dream of electronics engineers."
Patterson expects this technology - in a copper wire version - to be part of the AYK-14 avionics computer in such Navy Department applications as the Marine Corps Boeing AV-8B jump jet and the F/A-18 E/F aircraft. He says FDDI (also in copper) may go on the Marine Advanced Amphibious Assault Vehicle, although that, too, could go to a copper version of Fibre Channel.
Patterson says designers are still afraid of working with optical fiber because of problems with breakage and termination. Although Marine Corps experts actually did pioneer the use of a very short optical fiber databus in non-mission-critical functions of the AV-8B 20 years ago, the idea never caught on. Nonetheless, Patterson, too, says he sees opportunities for fiber optics in the JSF, which he says is "between a rock and a hard place" in its data needs to push the envelope of performance.
Overlooked in this issue of competing databus technologies, Patterson says, is the role of the asynchronous transfer mode (ATM) protocol, which is media-independent and in his estimation "flies on Fibre Channel." The bottom line, he says, is that, as prices go down the traditional experience curve, ruggedized Fibre Channel will make inroads on 1553 and, one by one, the databus producers will "join the scrimmage."
Championing Fibre Channel is Jack Staub, president of Delphi Engineering Group in Costa Mesa, Calif., who estimates this market at $600 million in 1997 and doubling this year. "This will be a $10 billion industry by the year 2000," he says, "and 40 companies are chomping at the bit to be the fastest and cheapest." He likens the situation to personal computers. "PCs started out like little toys," he says, "and now they`re workstations and they`re taking over the world."
Bandwidth is a logical driver, Staub says, but there are other considerations. In high-rate data applications such as sonar, designers in the past have put their digital signal processors as close as possible to the sensor to convert the data into small manageable chunks before processing. With increased bandwidth available to them, systems designers can place the converters kilometers away from the sensors. This optimizes the processing task, but Staub points out that it also permits sensors such as radars to be placed on what he calls the "awkward positions of an aircraft," such as the belly and wings.
He envisions a concept of real-time radar signal acquisition that is Fibre Channel-compatible. In this way it could use a COTS data recorder that would plug into the network. This would be particularly helpful for sensors that inherently require high bandwidth, such as focal plane arrays. The high bandwidth potential of Fibre Channel improves the links among processors, displays, sensors, and other subsystems. It also provides embedded sensor processing between processors and within the units.
"Fibre Channel will become the communications fabric of choice for embedded systems," Staub says. He lists six reasons why: cost-effective, non-proprietary, homogeneous interface, scaleable, low latency, and support of remote location of elements.
The venerable MIL-STD-1553 databus is entrenched in a wide variety of applications, yet the far greater bandwidth that Fibre Channel offers is sending many systems designers toward the newer technology
The Summit line of 1553 products from UTMC are targeted at smart munitions applications.
The BU-61588 5 volt miniature Advanved Communication Engine (Mini-ACE) from ILC Data Device Corp. is targted at small, cost-sensitive applications.
The Multiplex Bus Switch from Systran Corp. is a 16-6y-16 switch for use in avionics laboratories that develop operational 1553 systems.
Ethernet edges 1553 and FDDI in research organization survey
SUDBURY, Mass. - Ethernet was the top choice for networking communications among military users in a survey by Technology Research Institute (TRI) in Sudbury, Mass. Following Ethernet, the study lists MIL-STD 1553, FDDI, and Fibre Channel in that order. The survey was to elicit opinions on what computer functions users required in their systems.
Citing Ethernet were 70 of the 102 respondents overall and, unsurprisingly, topped the ground applications with 79 responses and naval applications with 90. For airborne applications, it trailed 1553, 68 to 55 responses.
The survey also provides insights into acceptance of Fibre Channel: 30 responses for airborne, 21 for naval applications, and only 9 for the ground. FDDI was the second choice behind Ethernet for naval applications and was also strong on the ground and in the air: 42 and 20 responses, respectively.
The survey also looked at such mature technologies as the Naval Tactical Data (NTDS) and found that, after 30 years, it remains entrenched. "Today, NTDS is such an integral part of the Navy weapons infrastructure that Navy officials have no choice but to keep using NTDS on the newest AEGIS class destroyers," according to TRI analysts. "The problem? Too many older naval gun and torpedo systems in the fleet communicate using mil spec NTDS."
While this may cause problems for Navy operational personnel, it keeps a market alive for the suppliers. TRI notes that "the demand for NTDS boards can only increase." Moreover, according to the survey, "This situation bodes well for NTDS board suppliers VisiCom in San Diego, GET Engineering in El Cajon, Calif., and Sabtech in San Diego.
Regarding 1553, TRI analysts say it has successfully made the transition from its avionics origin to become a standard for shipboard applications. "Today, it is used widely on a variety of military platforms, including missile systems and electronic pods carried on aircraft, ground vetronics, and shipboard and submarine weapon systems," the organization notes.
It still excels at securely transferring brief messages from one computer to another and provides a dependable control mechanism, TRI adds, but it suffers from its bandwidth limitations. "To boost throughput, a common solution is to keep adding more 1553 boards to a design," TRI concludes. - J.R.
Vista Controls launches Fibre Channel interface
SANTA CLARITA, Calif. - One of the first companies out of the Fibre Channel starting gate is Vista Controls Corp. in Santa Clarita, Calif.
Vista has introduced a conduction-cooled Fibre Channel PMC interface, the FC-PMC, with a sustained transfer rate of 102.7 megabytes per second. That is 99.7 percent of Fibre Channel`s theoretical maximum performance, Vista experts claim.
Vista Controls is working with Delphi Engineering Group in Costa Mesa, Calif. on the rugged Fibre Channel PMC project. Gorky Chin, vice president of advanced technology at Vista, says this is the first of several products to emerge from this partnership.
Downstream Chin says he is looking for future versions that will enable transparent migration from the company`s current data-buses to the 1 gigabit Fibre Channel network with equal capabilities, including fault-tolerant networks. Other planned enhancements include adding node bypass relays, a second fully functional Fibre Channel interface, and built-in-test and self-test capabilities.
The new product is essentially a ruggedized version of a commercial air-cooled development board to a conduction-cooled unit operating at -40 degrees Celsius to 75 C. Single-quantity pricing is $2,495 and delivery of development boards is 30 days after receipt of order. Further information is available on the company`s Web site, www.VISTAcc.com.- J.R.
The FC-PMC from Vista Controls is one of the first conduction-cooled Fibre Channel PMC interfaces to enter the market.
Fiber optic databus set for U.S., British Scout vehicle
WASHINGTON - "We will use fiber optics" in the databus of the Future Scout and Cavalry System (FSCS), declared Ken McGinty, managing director of SIKA International Ltd., one of two teams developing the vehicle for use by U.S. and British forces. SIKA is the partnership of Lockheed Martin Electronics & Missiles in Orlando, Fla., and British Aerospace in Farnborough, England.
Speaking at a briefing at last month`s Association of the U.S. Army (AUSA) conference in Washington, McGinty described the system architecture for the vehicle`s command, control, communications, computers, and intelligence (C4I) as the largest technical task and therefore the one receiving the most development attention.
Acceptance of a new technology for the databus is in contrast to the other facets of the vehicle`s vetronics, in which SIKA experts are trying to borrow technology from such other programs as the U.S. Army Boeing-Sikorsky RAH-66 Comanche scout/attack helicopter, the U.S. Army General Dynamics M1A2 Abrams main battle tank, and the British Army Vickers Challenger 2 main battle tank.
The development team has already achieved a 155-megabit-per-second rate, adds Julian Browne, strategic marketing manager for London-based Racal, a member of the SIKA team. Browne says he expects that to proceed upward to 622 megabits per second and then to one gigabit using Fibre Channel.
Driving the databus will be the asynchronous transfer mode (ATM), Browne says, and it will use commercial switching techniques. One of the considerations in the design is what he calls "the primacy of voice command" used with the data.
The SIKA team lists the common core C4I architecture and software as the "centerpiece" in its development, and the goal is to achieve plug and play plus a modular design to accommodate different reconnaissance missions. The program also works within a weight constraint of about 21.5 tons per vehicle to be transportable by C-130 aircraft. In addition to the HTI approach, the team is also looking at commercial powerpack options.
FSCS is part of the two armies` Tactical Reconnaissance Armored Combat Equipment Requirement effort known as TRACER, in which the Lockheed Martin-British Aerospace team is competing with another team headed by United Defense LP. The downselect for the engineering and manufacturing development phase is scheduled for 2002. - J.R.