Military Embedded Systems

Avionics databus users demand more reliability and flexibility

Story

May 16, 2024

John McHale

Editorial Director

Military Embedded Systems

Photo by Air Force Master Sgt. Michael Crane

MIL-STD-1553 designers are still winning new contracts for their products based on the venerable standard. Meanwhile, faster protocols such as Ethernet, ARINC 429, and Fibre channel also play important roles in military systems.

It’s been five decades since MIL-STD-1553 – the one-megabit-per-second databus standard – came on the scene, a contrast with today’s military aviation platforms, which leverage high-speed avionics databus protocols such as high-speed Ethernet, ARINC 429, and even Fibre Channel. Yet MIL-STD-1553 designs continue to be found in new systems and sustaining older systems, despite its slower speed.

All these standards, while varying in speed, offer what military designers want: robust reliability and determinism for mission-critical applications such as avionics, radar, electronic warfare (EW), and sensor processing.

“For data-intensive systems like radar, electronic warfare, signals intelligence (SIGINT), and sensor processing, military customers must have very efficient high-bandwidth networking that doesn’t tax the application’s processing elements,” says Greg Bolstad, Chief Technical Officer, Critical I/O (Irvine, California). “They need 1/10/25/40+ Gbit networks with low latency and highly deterministic behavior.”

Robust determinism is critical. “Key factors [we see] in component selection among military customers are reliability, determinism – especially in mission-critical systems – SwaP-C [size, weight, power, and cost], and interoperability over different systems,” says Anthony Murray, Director of Marketing for Holt Integrated Circuits (Aliso Viejo, California). There is also a growing use of commercial off-the-shelf (COTS) components rather than custom solutions if possible as well as a demand for ease of use, he adds.

Open architectures and interoperability are also important. “We believe that customers are looking for proven solutions that operate reliably to minimize their project risk,” says Aaron Bonner, Chief Engineer, Avionics Data, at Astronics Electronic Systems (Kirkland, Washington). “They want open architecture hardware and software solutions that are prevalidated and easily integrate on their platform.”

Often it’s not one databus above all others, but rather a mix of different standards.

“For military applications usually there is a mix of buses required,” says Alex Ivchenko, Ph.D., Director, Engineering at United Electronic Industries (UEI – Norwood, Massachusetts). “[While] most data is driven on the MIL-STD-1553 bus, some applications require the IEEE 1394 bus. ARINC-429 is still in wide use, as well as different serial interfaces for interdevice communications.”

“They are looking for databuses to be handled by nodal access units (NAUs) used in aviation or by IO adapters (IOAs) used on ground vehicles, which then put the avionics data (and other IO) on a digital back bone (DBB) using an 802.3 Ethernet or Time Sensitive Networking (TSN) interface,” says Lino Massafra, Vice President, Sales and Marketing, at North Atlantic Industries (NAI – Bohemia, New York). NAI does this with its NIU3A, a small, rugged, low-power, self-contained processing and multifunction I/O system that is preconfigured with 24-CH programmable discrete I/O, 8-channel ARINC 429/575, 4-channel CANBus, and 2-channel MIL-STD-1553 functions. (Figure 1.)

[Figure 1 ǀ The NIU3A from NAI is a small, rugged, low-power, self-contained processing and multifunction I/O system.]

“Outside of legacy avionics buses like MIL-STD-1553 and ARINC 429, we can see customers are looking forward at Ethernet-based avionics bus architecture,” Ivchenko says. “The earlier ARINC-664 (AFDX) bus appeared to be destined for its replacement; however, it lacked true real-time response and was heavily reliant on the functionality of AFDX switches.”

AFDX is also popular in commercial applications “as the main ‘backbone’ databus in Airbus A350, A380, A400, and Boeing 787 Dreamliner,” Murray says. For databus needs, Holt offers the 3.3V HI-2130, a BC/MT/RT MIL-STD-1553 terminal that integrates digital protocol, shared RAM with ECC, analog transceivers, and passive transformers in a single 15 mm by 15 mm compact package.

Ethernet is everywhere and ubiquitous thanks to its speed, interoperability, and availability. “Ethernet has become a recognized standard for avionics system backbones due to speed, bandwidth, and prevalence,” Bonner says. “ARINC 429, Serial, CANBus, and Enhanced Bit Rate MIL-STD-1553 (EBR-1553) are proven in use and are primarily complementary to MIL-STD-1553.” Astronics offers the Ballard NG1 Series Avionics I/O Converter, which simplifies the expansion and integration of avionics I/O with mission computing systems. It is easy to locate close to data sources and includes essential I/O such as MIL-STD-1553, ARINC 429, serial, CANBus, and discrete. (Figure 2.)

[Figure 2 ǀ Pictured is the Astronics Ballard NG1 Series Avionics I/O Converter.]

“In practice, adding new avionics databus solutions tends to not displace existing databuses as much as they highlight the need for effective and reliable protocol-conversion products,” Bonner notes.

Cybersecurity aspects

Cybersecurity is not something immediately thought of when discussing databuses, but it is becoming very important to the U.S. Department of Defense (DoD).

“Avionics databus solutions require a greater level of cybersecurity than ever before,” says Marc Foster, Western Regional Sales Manager, Sealevel Systems, Inc. (Liberty, South Carolina.)

“There are some flow-down requirements from the DoD with a stage set of requirements,” says Mike Hegarty, Marketing Manager at Data Device Corp. (DDC – Bohemia, New York) says. “There’s a big push for something called zero trust, which is an overarching approach to how you do cybersecurity. There is a roadmap for it on the IT side, but when you’ve got something like MIL-STD-1553 or ARINC 429 on an aircraft that doesn’t have traditional network stacks, it’s hard to apply rules and policies that come out of an IT infrastructure.

DDC researched cybersecurity related to MIL-STD-1553 and “what we can do to try and mitigate some of the cyber vulnerabilities that may exist when you start thinking about malicious implants, what can impact the system, and what can we do within the protocol layer to try and mitigate some of those types of attacks. We’ve got prototype hardware that we’re sampling our customers now.”

Fo cybersecurity, DDC engineers designed a cyber-resilient MIL-STD-1553 component called Total-ACE CR that is a drop-in replacement for DDC’s BU-64863T8 Total-ACE. It has a 1553 protocol interface layer, integrated memory, dual transceivers, transformers, and a Hardware Watchdog that mitigates inherent cyber weaknesses within the MIL-STD-1553 protocol.

Cybersecurity is at the front of everyone’s mind, but “not something that there’s an easy answer for as the best approaches for cybersecurity tend to be layered,” he says. Defense in depth is the idea that you’re going to have one magic firewall that’s going to solve all your problems, that you need to have multiple layers, Hegarty continues. “That’s what zero trust is about.”

MIL-STD-1553’s popularity

As someone once wrote, nothing is guaranteed except “death, taxes, and MIL-STD-1553.” So why is this slow-speed standard still popular in a high-speed world?

“MIL-STD-1553 is winning new contracts – it’s installed on so many platforms with so many sensors using MIL-STD-1553 that it’s not going away anytime soon,” NAI’s Massafra says. “However, they are starting to remove it from main processor LRUs and using the DDB approach.”

Hegarty calls MIL-STD-1553 “the classic definition of a dominant market standard. People use it because they’ve used it before. They use it because it’s in the aircraft, it’s part of the infrastructure. The idea is that as you start putting new electronics and updated systems onto aircraft, you’re not going to want to rewire the aircraft or if you do, you want to minimize it. Because the aircraft systems were all architected around MIL-STD-1553, it becomes the nervous system for the aircraft.

“And the reality is, it still works,” he continues. “It’s a highly engineered solution. It’s reliable. It’s robust and from an electrical environmental point of view, it does very well. So, in a lot of cases, there’s no reason to change it.

Part of the reason MIL-STD-1553 remains prevalent “is because it is a proven solution that has a history of reliable performance,” Bonner says. “The deterministic aspects of the protocol are time-tested and effective at accomplishing the needed tasks. It has high noise immunity that is not replicated well in other alternatives, and low RF emissions.”

“It is a well-defined bus with real-time performance and a critical mass of available LRUs. Lower data rate is usually not a problem for delivering in-flight data,” UEI’s Ivchenko notes. “Where it becomes a problem [lies] in significantly slowing down the ability to upload maps and other high-bandwidth data in preparation of missions and maintenance.” UEI offers the DNx-1553-553, a high-performance, two-channel MIL-STD-1553 card that is a MIL-STD-1553-to-Ethernet converter. (Figure 3.)

[Figure 3ǀ The DNx-1553-553 from UEI is a high-performance, two-channel MIL-STD-1553 card that is a MIL-STD-1553-to-Ethernet converter.]

Strong supply chains also enable longevity. “MIL-STD-1553 has a broad, well-established base with a large service history and is used widely on most if not all platforms,” Murray says. A deterrent for moving away from MIL-STD-1553 is the large base of existing software applications associated with MIL-STD-1553 and the cost of redesign and requalification, he notes.

Military applications desire these characteristics, especially command-and-control applications, which require an application that is “robust, deterministic, fault-tolerant and dual-redundant with a transformer coupled bus,” Murray continues. “Its variant, MIL-STD-1760, is widely used in mission computers and weapons platforms.”

Reliable standards also enable connections between legacy and modern technology: “While new technologies are always emerging and providing more advanced capabilities, these advances in technology also take time to mature,” Foster says. “To bridge that gap, existing technologies like MIL-STD-1553 allow new platforms to utilize existing subsystems on new platforms. [It also extends] the life cycle of existing platforms like the F15 and F16 and their support equipment for the same reasons.”

Fibre Channel still in play

While not quite as venerable as MIL-STD-1553, Fibre Channel is not a new standard, but is also still going strong. “Fibre Channel established itself 20+ years ago as the most efficient high-performance network choice, well before many of the enhancements making Ethernet a suitable choice for the same type of systems,” Critical I/O’s Bolstad says. “That helped make Fibre Channel the de facto choice because it was up to the job well before other choices suitable options were available.”

Speeding up MIL-STD-1553

Over the decades, many in the engineering community thought they could speed up the MIL-STD-1553 standard, but not many of these attempts took hold. There are still efforts going on in this arena, but the availability of other and faster choices make some feel it’s not necessary.

“There have been several attempts at increasing the speed of MIL-STD-1553,” Bonner notes. “At this point, none have seemed to gain the wide acceptance needed to displace MIL-STD-1553. Higher-bandwidth applications have often moved to other data buses such as Fibre Channel or Ethernet, reducing the need to produce a faster MIL-STD-1553 variant.”

There has always been a demand for more speed and always will be. “There’ve been a lot of concepts for speeding up MIL-STD-1553,” Hegarty says. “We did a fair amount of research into that technology. We basically showed that yes, it’s possible to speed up MIL-STD-1553 and get to a 10, maybe 100-time increase in the data rate.”

The question becomes, Hegarty asks, what is the cost? And then what’s the benefit? “So, the idea is what can you do at 10, 50, maybe 100 megabits per second that you can’t do with one megabit per second? And if that’s not going to give you some breakthrough capability, why invest in it? When we start looking at different needs that are out there and what people are looking for, they’re not looking for 10 megabit or 100 megabit, they’re looking for 10 Gig going to 100 Gig as the emerging requirements for sensors and displays and videos need very high-speed interfaces.”

“A higher-speed example that has seen success is EBR-1553, which communicates at 10 Mb/sec and can be utilized with applications like rapid reprogramming of smart munitions,” Foster says. “EBR-1553 utilizes a star topology rather than the multidrop topology of standard 1553. For applications that need higher speed but need MIL-STD-1553’s familiar multi-drop bus topology, there are solutions that offer MIL-STD-1553 with speeds of 5 Mb/sec that maintain the time-proven resiliency of standard MIL-STD-1553.

Looking forward – TSN

Time Sensitive Networking (TSN), a more deterministic version of Ethernet, is gaining attention in new military requirements.

“For higher-level interconnection of subsystems and sensors, we will see increasing reliance on 10 Gb to 100 Gb Ethernet, leveraging recent developments in Ethernet such as Precision Time Protocol and [TSN],” Bolstad says. Specifically, DDC sees TSN as a fundamental building block for high-speed data paths for future platforms, Hegarty says. “Several major [military] programs out there have selected or are looking at TSN.”

In addition to internal TSN development, DDC has been involved with an industry standard being developed under a joint collaboration between the Society of Automotive Engineers (SAE) and the IEEE 802.1 Working Group, which is basically the standards body for Ethernet, he continues. [Editor’s note: The TSN Task Group is a part of the IEEE 802.1 Working Group (WG).] “The group is tasked with coming up with a profile of TSN for aerospace networks,” Hegarty adds.

“These and other recent additions make Ethernet much more suitable to real-time applications by providing time synchronization, along with guaranteed bandwidth, latency, and delivery for time-critical data,” Bolstad says. “Similar to Ethernet at the physical level, but using a much different protocol, the use of InfiniBand will also likely increase, specifically for large multichassis HPC-like processing architectures. And finally, well-established protocols such as MIL-STD-1553 and Fibre Channel will continue to be relevant for many years due to their broad usage and proven performance and reliability in many existing military platforms.”