AUSA 2019: Product Roundup
The Association of the United States Army (AUSA) held its annual meeting and exposition in October, where vendors get a chance to showcase their products and technologies for the military market. This year’s conference highlights a range of technologies from development systems and servers with artificial intelligence (AI) engines to multi-function displays and connectors. Whether the product is a system or a component, vendors are focused on meeting the latest requirements in the areas of modern warfare, cybersecurity, enhanced networking capabilities and communications, and autonomous mobility.
Here’s a sampling of products showcased at the exposition.
General Micro Systems announced that its rugged vehicle-mounted server combo and AI engine, suitable for massive computation and sensor fusion in autonomous vehicles, unmanned aerial systems (UAS), and C4ISR/electronic warfare systems, was selected for next-generation army vehicle and airborne systems. Delivered in two small “shoebox-sized” rugged chassis, the S422-SW and X422 server combo offers lots of server processing power, 10/40/100 Gigabit networking ports for sensors, and general-purpose graphics processing unit (GP-GPU) AI.
The two programs that selected the S422-SW “Thunder” and X422 “Lightning” combo will deploy it in mobile platforms to move IP-based sensor data instantaneously over multi-sensor LANs into the server and AI processor. Once processed, the server provides information that can help maneuver a vehicle or UAS in real-time, calculate a fire control solution for a weapon, or identify threats such as stationary IEDs or incoming objects such as projectiles, according to General Micro Systems.
Using commercial-off-the-shelf (COTS) components, the systems include technologies from Intel, Nvidia, Broadcom, and Cumulus Networks. The S422-SW is a conduction-cooled, fan-less, rugged, low-cost Intel Xeon E5 server operating over the temperature range of -40°C to 85°C. It provides an on-platform or in-vehicle 30-port 10 Gigabit Ethernet local area network (LAN) designed to interface with high-bandwidth sensors needed for next-generation autonomous vehicles or battlefield reconnaissance.
Sensor data is moved in real-time over the LAN, processed, and stored locally by the S422-SW’s Intel Xeon E5 server CPU. The companion X422 co-processor uses two of Nvidia’s V100 Tesla GP-GPU AI engines to search the data to perform tasks, such as target tracking, image processing and enhancement, and vector algorithms, in real time at 400 FLOPS.
“Thunder” is also an enterprise-class multi-port LAN or a network attached storage (NAS) appliance equipped with a professional-class intelligent Layer 2/3/pseudo-4 Ethernet switch/router and data center networking software from Cumulus Networks. Networking capability includes four 40 GbE fiber ports and thirty 10 GbE ports. The 10 GbE ports come from a Broadcom Layer 2/3/pseudo-4 enterprise-class switch that has never before been used in a deployed battlefield computer, according to the company.
Each of the 10 GbE ports support power-over-Ethernet (POE+) to directly power remote nodes or sensors, while simplifying wiring requirements, up to 100 W maximum total power. The quad 40 GbE fiber ports — also configurable as 100 GbE— are used for the highest-rate sensors, or inter-system communications to SDRs, data-intensive EW processors, or other in-vehicle systems.
The S422-SW works with the X422 via the GMS FlexIO 16 lane, 8 GT/s PCI Express Gen 3 bus extension. The X422 uses the dual Nvidia Tesla V100 data mining/algorithm processing AI engines that deliver a 400 TFLOPS of algorithm computation. The GP-GPU modules are ruggedized by GMS for reliable conduction cooling in the X422 chassis. Other dual-slot, 250-W PCIe cards can be used for co-processing with the S422-SW, including AMD GP-GPUs, Altera or Xilinx FPGA modules, and DSP cards. Local intelligence in X422 allows the cards to work together or independently, depending upon the application.
General Micro Systems (GMS) also claimed the industry’s first sealed, fan-less, conduction-cooled rack-mount servers with AI, featuring mil-circular (38999) connectors for ruggedness in demanding defense and aerospace applications. The American-sourced and manufactured fully configurable server, TITAN, is available in a 1U or 2U chassis and uses up to four of Intel’s latest 2nd generation Scalable Xeon processors. An air-cooled version also is available.
Key specs include:
- Up to two Nvidia Tesla V100 enterprise-class GPGPU co-processors for AI workloads—offering data center performance on the battlefield (up to 400 TFLOPS)
- A first-of-a-kind removable secure high-speed storage (SHS) canister with up to 64 TB of RAID-encrypted storage for data recorders, sensor data storage, or system de-classification
- 1/10/40Gb Ethernet built in with copper or fiber ports, plus available embedded 20-port switch — via the 38999 mil-circular connectors or standard COTS connectors
- Up to 10 PCIe Gen 3 add-in slots for program-specific functionality, I/O, or custom functions
- Dual-redundant 110/220 VAC 60 Hz/400 Hz power supplies for each CPU pair; MIL-STD-704 with 50 ms hold-up recovers from “dirty” power found in typical military systems
- Dedicated out-of-band remote management via IPMI
- GMS SourceSafe BIOS and SecureDNA firmware/BIOS/storage sanitization
Read more of the AUSA 2019 Product Roundup here.
November 5, 2019