Deterministic Ethernet in military front lines

Increasingly, embedded edge systems are used to collect, analyse and communicate data from the field in real-time — driving a need for deterministic network protocols. Here, Martin Frederiksen, Managing Director of Recab UK, has explained the importance of deterministic Ethernet in military operations.

Whether they are onboard military drones used for reconnaissance or on naval warships to communicate their position, embedded computing systems are widely used in the military sector. This is only set to continue, as growth predictions from analyst firm MarketsAndMarkets anticipate the global military embedded systems market size to be valued at $18.4bn by 2025, growing at a compound annual growth rate (CAGR) of 6.4%.

The growth is impressive but not without its challenges. Military edge embedded systems must not only meet strict regulatory standards, but they must also be able to exceed demanding operational expectations.

Distributed military computing must be able to offer low-latency, high-availability, interoperability between networked devices and high-bandwidth. Many applications, such as radar, electronic warfare, signals intelligence or weapon control systems, have extremely tight latency constraints and demand accelerated edge computing via field-programmable gate array (FPGA) devices. In most cases, networked applications need to be able to synchronise within a nanosecond time range.

Through Recab’s extensive work with military and defence system developers, we have noticed a clear trend towards Ethernet as the network protocol of choice due to a relatively low cost and ubiquitous nature. However, Ethernet is generally non-deterministic — something that makes it typically reliable but impractical for mission-critical, dynamic systems.

Deterministic Ethernet

There are protocols that can be used to support Ethernet in meeting mission-critical system demands. As an example, to benefit from zero-delay recovery time and no-frame loss in the event of network failure, the sector is adopting Ethernet-based protocols like high-availability seamless redundancy (HSR) and parallel redundancy protocol (PRP), both outlined in IEC 62439-3.

HSR provides redundancy by sending packets in both directions through a ring network. A simple HSR network consists in doubly attached bridging nodes, each having two Ethernet ports. Both ports send the same data frame so that, in a fault-free scenario, nodes receive two identical frames. This means that even if the fibreoptic is broken, no frame loss is ensured and the communication among all network nodes continues.

PRP, on the other hand, implements redundancy in the nodes rather than in the network. dual attached nodes (DANs) are connected to two independent and standard Ethernet networks (LAN A and LAN B) and send the same frames over both networks. The PRP operative ensures the receipt of all the information, even if one network fails.

For standardised deterministic Ethernet networking, the best option is time-sensitive networking (TSN). TSN is the new generation Ethernet that provides determinism to enforce advanced quality-of-service policies. It supports in the same network hard real-time, reserved and best-effort traffic. The seamless redundancy at frame level is supported by the IEEE 802.1CB TSN standard.

Simplifying connectivity

With the continued move to real-time, high-bandwidth and open networks of fixed and dynamic embedded systems, the process of determining a scalable system naturally becomes more complex. Fortunately, Recab UK has been applying its extensive industry experience to help military engineers overcome this complexity. One way we’re able to do this is by partnering with military network computing specialists such as SoC-e, whose Relyum range offers an ideal all-in-one solution.

Relyum by SoC-e is a range of military commercial off-the-shelf (COTS) managed 1/10G ethernet switch, router and edge computing equipment. It supports up to 20x 1G copper and up to 6x 1/10G fibreoptic ports, with support for different media types and its distribution in the MIL-DTL-38999 connectors allowing for complete and cost-effective network infrastructures.

The product specifically boasts a Xilinx Ultrascale+ MPSoC device that includes six ARM CPUs, one GPU and a FPGA in the same integrated circuit. The switching and routing functions are accelerated by hardware in the FPGA section, which gives it the flexibility to meet various military requirements. This flexibility allows the Relyum to support HSR or TSN protocols with the same hardware. Fibreoptic rings combining these protocols are feasible thanks to the inter-switch coordination mechanism developed by SoC-e.

The flexibility makes it an ideal option for numerous military applications, whether on land, in air or at sea. For example, military land vehicles benefit from zero-delay recovery time, deterministic Ethernet communication, and can also take advantage of the equipment to embed hardware-software microservices oriented to sensor data pre-processing. The switches are also Generic Vehicle Architecture (GVA) compliant, making them an ideal fit for defence vehicles.

Similarly, naval warships that combine different timing sources and protocols, like NTP-PTP-GNSS, with cyber-security and multi-media networking interfaces benefit from the flexible concept of Relyum by SoC-e. Recab can work with military OEMs to develop a custom FPGA and software design, combining a wide portfolio of IP cores for networking, synchronisation and security.

As the use of dynamic, edge-embedded systems in military settings continues to increase, more military OEMs will experience mounting complexity when establishing networks. Ensuring that data can be shared as efficiently and quickly as possible, in an environment where speed is of the essence, is imperative — but it does not need to be so complicated that it slows down project development.