Innovation Stories

Concurrent Engineering

Complex engineering development processes often result in long design cycles, sometimes measured in years. And that simply won’t do. Warfighter needs demand a far more rapid development and deployment of embedded software.

Thanks to the growing use of Intel processors in defense platforms, a much better methodology exits: concurrent engineering. Concurrent engineering enables both software and hardware development to happen in parallel, resulting in faster time to deployment, reduced risk and protected software investments.

In addition to the usage of Intel processors, Mercury also offers a unique “erector set” lab target hardware chassis. This flexible unit allows you to perform final application debug along with performance tuning and regression testing.


For more information on our concurrent engineering methodology, download our whitepaper that explores the 5-step process we have developed for accelerating the deployment of embedded subsystems.

Electronic Countermeasures

Complex engineering development processes often result in long design cycles, sometimes measured in years. And that simply won’t do. Warfighter needs demand a far more rapid development and deployment of embedded software.

Protecting our platforms and warfighters from radar-directed attack is the new front in modern warfare. And that requires flawless jamming and deception. Mercury’s digital RF memory (DRFM) subsystems enable signals to be detected, captured, modified and returned with extremely low latency, all from a deployed platform. Using commercially available components like FPGAs, our DRFMs fit inside a rugged, compact enclosure to meet mission-specific requirements.


To further explore the role of DRFMs in modern electronic countermeasures, download our whitepaper, "Leveraging Digital RF Memory Electronic Jammers on Modern Deceptive Electronic Attack Systems".


Field programmable gate arrays (FPGAs) offer many benefits for the defense environment, including high-performance and “on-the-fly” re-programmability. However, FPGA IP (typically VHDL or Verilog code) is difficult to develop because it requires detailed hardware and system integration, as well as control knowledge.

Mercury offers several innovative tools to address this situation. Our OpenFPGA Developer’s Kit eases application development by supplying optimized infrastructure IP to abstract the lower-level hardware details and system utilities. The Simulation and Verification Environment (SVE) is a complete end-to-end simulation and verification tool for Mercury FPGA-based systems. It is designed to enable application developers to quickly model and verify FPGA code, dramatically reducing time to market. Finally, our Services and Systems Integration (SSI) team can create custom IP specifically designed for your environment.



Mercury has created a way to maximize GPGPU performance for sensor processing applications through an IP innovation called StreamDirect™.

Without affecting application software, StreamDirect moves data with great efficiency into GPGPU coprocessors, achieving performance improvement factors that approach 2X. Previously, systems that used GPGPUs had to first pass the data to a CPU’s memory and then transfer the same data from the CPU’s memory to the GPGPU. StreamDirect eliminates this cumbersome copy step, creating a direct, high-bandwidth DMA channel between the sensor and the GPGPU. Now applications such as EO/IR, radar, cyber and EW can benefit from faster intelligence.



Mercury has developed a method for ruggedizing the industry standard MXM connector for GPGUs, a packaging innovation that allows new generations of cutting-edge GPGPUs to be rapidly deployed in harsh environments.

With the ruggedized MXM connector, deployable GPGPU-based subsystems can be updated quickly with new, increasingly powerful GPGPUs. ISR applications can leverage these performance advantages driven by the commercial electronics market. Mercury’s modular GPGPU MXM designs enable prime contractor customers to preserve their software and IP investments while accelerating their program development schedules and reducing program risk.


High-Capacity Deployable Storage

In the defense industry, modern sensors are capable of generating massive amounts of real-time data that must be stored for in-depth analysis, future reference or other near-real-time needs. As a result, platforms carrying these sensors face particularly challenging storage design issues — especially given their unique SWaP considerations and their exposure to harsh temperatures, shock and vibration. Platforms that fall into this category include unmanned aircraft systems and ground mobile systems.

Mercury’s customers find our approach to deployable storage solutions highly innovative, because every storage solution is completely customizable to their specific needs. For example, our deployable storage units have been designed to specifications such data interfaces, storage capacity (up to 96 TB), redundancy and security requirements.  Also, the deployable storage units are based on commodity solid-state disks (SSDs) for a long-term availability of supply. This high degree of customization, flexible and modular design, combined with a clear upgrade path lets you future-proof your storage designs.


High-Power RF Capabilities

The ability to successfully monitor, manipulate and exploit enemy electronic transmissions is critical to success in the war on terror. The explosion of new communication technologies makes achieving that dominance a truly significant challenge. To counter these threats, the DoD’s goal is “full-spectrum dominance.”

The interception of both known and unknown signals is essential to achieve this goal. Mercury’s high-power RF products and systems provide a high probability of intercept (HPOI), even for interception of signals with very short duration. In fact, Mercury’s high-power RF capabilities power the leading HPOI systems on the planet.


Intel x86 Server-Class Processing

Mercury was the first company in the high-performance embedded computing industry to bring Intel® enterprise server-class processing to deployed sensor-based systems. The Ensemble 6000 Series OpenVPX Intel Xeon Dual Octal-Core HDS6601 Module delivers nearly half a teraflop in a single standard OpenVPX slot, more than four times what was possible with prior generation products.

Several significant Mercury innovations have made the ‘flying server’ a reality. Innovations in packaging and thermal management have made it possible to install a server-class processor (which generates copious heat) onto an embedded board. Another innovation called ‘standing memory’ allows for up to 32 GB of DRAM memory to reside on a 6U OpenVPX board. Finally, the POET (Protocol Offload Engine Technology) interconnect, through its low-latency and high-bandwidth characteristics, enables scalable, high-speed, and deterministic communications and I/O. And through its flexibility and versatility, POET interconnect efficiently extends the capabilities of the Intel-based subsystem.


OpenVPX Standard Development

Introduced in 2004, VITA 46 (VPX) was a high-performance and extensible embedded multicomputer specification. However, many of Mercury’s customers still expressed concerns over VPX system-level interoperability. To address VPX interoperability concerns, Mercury took the lead in forming the OpenVPX Industry Working Group. The working group set out to tackle several objectives, including reducing the time to market for developing COTS solutions in 3U and 6U form factors, increasing supplier options, lowering adoption risk with improved interoperability, reducing delivery times for new production systems, and reducing the total cost of system ownership.

This visionary effort immediately gained the support of other key COTS embedded suppliers, prime contractors and leading systems integrators. Their goal was to create and publish a top-down, system-level specification for module, backplane and development chassis architectures and pin-outs necessary to create interoperable VPX systems and assemblies. OpenVPX (otherwise known as VITA 65) was introduced in October 2009.

Mercury believes that the Open VPX System Specification with multi-plane reference architectures provides the basis for delivering innovative processing solutions designed to meet or exceed current and future DoD-critical, real-time embedded SWaP application requirements. 


Optimized Signal Processing Algorithms

Signal, image and data processing applications demand the greatest performance achievable. Mercury has developed, tested and optimized libraries of ultra-high performing, efficient, single- and multi-core processing libraries for the following Mercury-supported platforms: Freescale, Intel, NVIDIA and AMD. These algorithms are available in MultiCorePlus MathPack.

The algorithms contained in MultiCorePlus Mathpack automatically optimize for all cores on the supported processor. These algorithms let the developer focus on the behavior of their software application, rather than misunderstand complex mathematics of signal and image processing.

MultiCore MathPack represents the culmination of over 20 years of expertise in algorithm design and code optimization by Mercury’s staff of mathematicians, computer scientists and applications experts.


PowerPC Processing

Mercury has enjoyed a long-term association with Freescale Semiconductor. In fact, the innovative industry-standard RapidIO interconnect was developed by Freescale and Mercury.

For several years now, Mercury has been developing boards and systems based on AltiVec-enabled Power Architecture Freescale processors. These customers have protected their application software investment because their legacy Scientific Algorithm Library (SAL)-based applications can execute unchanged on the latest Power Architecture products, saving many hours of development time and minimizing development budget. Mercury’s latest Freescale-based product, the Ensemble 6000 Series OpenVPX HCD6210 Dual QorIQ T4240 Processing Module, greatly improves system-wide SWaP through a six-fold increase in processing capability per 6U slot versus previous generation Freescale-based products.


Real-Time Big Data Processing

Enormous amounts of data are generated daily by individuals, systems and our adversaries. Indeed, the current collection of data exceeds the capacity of traditional databases and software tools. Yet, the 21st century is a time of increasingly complex legal and data security needs. Clearly, legacy intelligence systems cannot meet 21st century needs. 

Mercury solves the Big Data problem by providing the best-of-breed solutions developed by skilled intelligence community experts to transform perishable data into persistent intelligence. Mercury’s high-performance, high-volume, real-time Big Data solutions perform data ingestion, analytics and storage on both structured and unstructured data. This not only enables the confidence in our highly sensitive data, but also protects it.



High-performance embedded computers are clearly valued because of their high speed of operation. When operating at high speeds, however, the modules generate large amounts of heat, potentially causing system damage.

Additionally, as technology continues to advance, unforeseen issues arise. Processing performance continues to increase year over year due to the availability of faster and faster processors, creating a significant demand to increase the interconnect speeds of embedded multicomputer systems. While InfiniBand®, 40 Gigabit Ethernet (40GbE) and PCIe® 3.0 possess the communication performance to address these needs, they have their own challenges when implemented in rugged, deployed systems. Mercury and TE Connectivity are collaborating to address these challenges and focus on next-generation solutions for robust module connections in high-vibration environments inherent in industrial and defense applications.


To learn more about the issues Mercury and TE Connectivity collaborated on and the rugged solution they developed, download "Rugged Technologies for Deployed Applications".

Thermal Management

Operation in demanding environments while maintaining requisite functionality and performance is a huge challenge for embedded applications. Mercury is addressing and even preempting these challenges. Innovations in ruggedization and packaging include the creation of new socketing technology and standing memory for Mercury’s line of Intel Server-Class Processing-based product line. Another innovation is the application of the industry-standard rugged MxM form factor for the GPGPU-based product line. 

Mercury applies many thermal management innovations at multiple points in the product development life-cycle. In the beginning of the design process, mechanical and thermal modeling is used to validate design approaches. Innovations in mechanical packaging support more efficient heat dissipation while maintaining the characteristics necessary to support high levels of shock and vibration. Modules are covered with various types of coatings to protect against high humidity, chemicals and fungus, and metal surfaces. Board thermal stresses are minimized through a number of design innovations, including technologies to support the mounting of high-powered mezzanine modules on compute baseboards.


For more information on Mercury’s Air Flow-By tm cooling system and how it can help increase performance while still meeting SWaP constraints, download "Innovations in Thermal Management: Air Flow-By".