Integrators save with preconfigured subsystems

Story

June 20, 2011

Duncan Young

GE Intelligent Platforms, Inc.

COTS application-ready subsystems live up to their names by keeping risk and time to deployment low and technology readiness high.

In general, an embeddable subsystem is a set of hardware and software modules working together to perform a specific function or set of functions needed by a specific platform or vehicle. Subsystems can be embedded within a larger system, for example an SBC plugged into a VPX (VITA 46) chassis, or a subsystem can be self-contained in its own chassis with a network or avionics interface connection to a larger system. Rugged SBCs are offered by many COTS vendors as “application ready” including preconfigured Built-In-Test (BIT) plus boot and recovery firmware and can be supplied with a preconfigured operating system such as VxWorks, Linux, or Windows with all the drivers needed to support the user’s application. The procurement of SBCs using the vendor’s specification and quality assurance systems is standard industry practice, but this concept has not been widely adopted at the higher level of an integrated chassis.

System integrator’s perspective

To be successful, a system integrator must reduce risk, demonstrate a high state of technology readiness, and reduce time to deployment whether procuring a component, a COTS SBC, or a subsystem. Integrators are encouraged to use COTS products in as many platforms as possible, yet the perception is that the Technology Readiness Level (TRL) of board-level COTS products is relatively low as they still require integration, packaging, and verification to fit the requirements of their target applications.

Compatibility and interoperability

While COTS vendors make every reasonable effort to ensure that their products are designed to common specifications, it is impossible to test every combination of hardware, packaging, and software during initial design verification. With a typical vendor’s product portfolio, this would lead to thousands of permutations for test. As a result, experience, best practice, and product life-cycle status are used to verify new designs. In some instances, to introduce an innovative concept, a standard might be adopted in an unusual way that is then not interoperable with every other similar product. An example of this is PMC/XMC, where power dissipation has grown from 15 W to 45 W, thus introducing special cases that can only become apparent in a small number of configurations. Similarly, VPX has encountered issues with control, data, and I/O plane allocations, prompting the introduction of OpenVPX (VITA 65) to create an assured level of system configurability.

Better business proposition

Chassis and power supplies also require configuration to allocate slots, interconnects, and external connections to suit an integrator’s specific requirements. The resulting unique configuration must then be fully documented, environmentally characterized, and functionally tested. However, three key technology forces are changing this complex business proposition:

PCI and PCI Express
Network architectures
SBC functionality and performance

PCI and PCI Express have introduced plug-and-play capability to system configuration; networks and Internet Protocol (IP) have revolutionized intra- and inter-subsystem communication. Finally, SBCs now have very high levels of performance and comprehensive application-oriented software support. A single SBC is well able to support even complex subsystems and, with the addition of PMC/XMC modules, can support a wide range of external I/O. The result is that many varieties of subsystems can be satisfied by a small number of variables in a chassis with only a few slots. This allows generic configuration, test, and characterization of a complete off-the-shelf, application-ready subsystem by the COTS vendor.

COTS subsystems

By reducing the variables, yet maintaining performance and functionality, a configured COTS subsystem is procurable off-the-shelf against its manufacturer’s specification with a TRL significantly higher than those of its individual components. An example of this is the CRS range of two- or three-slot rugged subsystems from GE Intelligent Platforms, depicted in Figure 1. They are 3U CompactPCI-based, preconfigured with either Freescale MPC7448 or Intel Core SBCs with combinations of ARINC, MIL-STD-1553B, CANbus, DAC, ADC, network, and serial interfaces.

Figure 1: A CRS rugged subsystem from GE Intelligent Platforms

(Click graphic to zoom by 1.9x)

Ideally suited to harsh environments, embedded computing applications in Unmanned Aerial Vehicles (UAVs), helicopters, small aircraft, and ground vehicles, an application-ready subsystem reduces risk and timescales to deployment for first- and second-tier systems integrators. In the future, further reductions in vital Size, Weight, and Power (SWaP) for these types of high-performance platforms can be anticipated, as the same principles of preconfiguration and verification are applied to, for example, newer architectures such as 3U VPX.

To learn more, e-mail Duncan at [email protected].