Point-to-point and packet-switching systems can co-exist in today’s cost-conscious military
“It’s like turning around an aircraft carrier.”
This analogy is often used in the business world to indicate how hard it can take a large entity to change direction, usually amounting to months or even years. But for the United States Navy, which is charged with turning these massive vessels, turning around an aircraft carrier pales in comparison with keeping communications networks up to date. Ironically, this complex undertaking fits the analogy perfectly.
When dealing with sensitive information pertaining to the nation’s defense, communications networks in the military must be highly secure and the infrastructure rugged enough to withstand the harshest environments. As the computer age developed in the twentieth century, Military Standards (MIL-STDs) and Military Specs (MIL-SPECs) ensured reliability and security. In the latter part of the century, however, the commercial sector’s computer advancements overtook the military’s capabilities. Commercial standards became part of the Navy’s computing infrastructure, and COTS (commercial off the shelf) equipment has since been housed in MIL-SPEC enclosures.1
Two distinct protocols
The Naval Tactical Data System (NTDS) was formed to link together computers and radars from multiple ships in order to form a more cohesive response to air threats. A new Military Standard was later developed, MIL-STD-1397, which define "the requirements for the physical, functional and electrical characteristics of a standard I/O data interface for digital data."
Today, the Navy uses multiple types of networking systems which are either based on NTDS, a point-to-point interface characterized by two specific end points and no data formatting, and Ethernet, which uses a packet-switched interface. The stark contrast in the nature of these protocols and the complex translation requirements makes networking between the systems a daunting challenge. To successfully get NTDS systems and Ethernet systems to communicate seamlessly, an intermediary is needed to handle the translation.
Bridge systems respond to the challenge
When upgrading computer systems, time is a precious commodity. Budget constraints often present engineers with a tradeoff between the speed and other benefits of a COTS system at a premium cost while having that same equipment communicate with peripheral equipment whose replacement is being deferred. Developers can, however, turn to specialized converters that can reside with the replacement system and can perform the complex data translation from NTDS to Ethernet without altering the original system.
These converters come in rackmount units for laboratory use and box form for use in harsh environments. The heart of NTDS to Ethernet translation devices is the FPGA (field programmable gate array). General purpose processors cannot handle the very fast NTDS handshake signals that take place as the bits go from the NTDS interface and are packetized for transmission across the LAN/WAN. FPGAs have the parallel processing capabilities necessary to respond to signals in hundreds of nanoseconds, and they can be customized to accommodate legacy equipment that does not comply with the NTDS specification. The PowerNet family of bridge devices from Sabtech, for example, includes firmware that packetizes the data and sends it across the Ethernet.
Meet future needs without an upgrade in sight
High speed, however, is just one of the characteristics that communications equipment vendors seek when choosing an FPGA. They must balance military communications systems’ need to last and simultaneous adapt for the future, all while COTS computer devices continue to evolve very quickly. “We want an FPGA that has low power consumption and sufficiently large memory capacity,” says Bryan Ly, senior software engineer at Sabtech. “Once we place all the design and logic – the code within the FPGA – there’s still at least 20 percent left over so that there’s room for expansion.” Additionally, they expect an FPGA, like the entire bridge system, to stand up to the rigor of rugged military conditions such as extremely high and low temperatures, dusty environments and high impact.
Like the devices they are trying to bridge, NTDS to Ethernet systems vary widely among vendors. Therefore, the first requirement for designers is to know the systems they are incorporating into the infrastructure. NTDS, for instance, has several iterations – both for parallel and serial interfaces – depending on voltage, transmission rate, signal type and other variables. So, the bridge system must accommodate the devices’ specific standard as defined within MIL-STD 1397. The Ethernet side, meanwhile, works on a standard TCP/-IP network stack. This is where the routing from point-to-point to general networks takes place, and multiple endpoints can be connected without physically switching cables. When selecting an appropriate transport protocol to use over Ethernet, designers must understand the tradeoffs between TCP and UDP. For example, while TCP offers reliable data transfer, it would not be appropriate on systems with low latency requirements. Designers can find bridge systems that balance these requirements through redundant network connections, where duplicate packets are sent out on both Ethernet interfaces. In the event of a lost packet, the duplicate packet may still reach the destination in time.
Bridge systems are intended to extend communications between legacy and newer systems, well into the future. The engineer must ensure that the bridge system supports not only the operating systems of the equipment they are currently trying to link, but also devices that may be added to a system later on. Designers should ask themselves whether the systems will support only Windows, or will they need a system that supports a large number of operating systems. When integrating software, it becomes critical to select a bridge that offers an operating system agnostic API. As Ly notes, “By writing code that works for one operating system, you can leverage that existing code and easily port it to another operating system with little to no change to reduce the risk of migration and reduce the amount of design time required.”
Of course, these considerations matter little unless the translation system itself is built for long life and provides security features. That means, a device must stand up to EQT testing where it absorbs blows from a 400-lb. hammer in three different axes nine times, among other types of punishment. Leading vendors also offer bridge modules with optional fiber-based interfaces that can help connect legacy systems over Ethernet, increasing their communications range over several kilometers – well beyond the NTDS maximum distances – while eliminating the resistance and impedance issues inherent with copper wires. So, for instance, a radar system and a command and control system can be situated miles apart without putting humans in harm’s way should something go wrong. Security features may be required, so consider the bridge system’s encryption options for data transfers, access control and remote configuration.
As Navy systems are pushed beyond their expected lifetimes, engineers struggle to integrate their point-to-point interface-based systems with packet-switched systems to maintain a modern communications infrastructure. Bridge systems give designers the flexibility they need to glean the benefits of modern communications systems without performing the sometimes prohibitively expensive u-turn of a full, COTS upgrade.
1 Michael Carter and Don Anderson., “Developing COTS Systems to Replace Legacy Systems,” www.ecnmag.com (http://bit.ly/187ltI0)