congatec is a high-performance organization and a global leader in embedded computing technologies. Christian Eder, Director Marketing at congatec discussed here more about the platform, product advantages and market growth.
Can you please discuss about high performance embedded computing platform and its support?
These days it is the edge that is driving the needs in high performance embedded computing (HPEC). Cloud applications are transforming to redundant and real-time capable fog servers close to the applications and often require a dedicated edge server setup. The most efficient way to design tailored, application-specific edge servers is by utilizing Server-on-Modules as they deliver ultra-scalability beyond the processor socket as well as highly cost efficient customization and performance upgrades for the next generation. Up until now, COM Express has been the only open Server-on-Module standard supporting HPEC via its Type 7 pin out. And it offers already a lot: Within the Intel family, the performance is ranging from Intel® Atom C3000 processors with 16 cores up to Intel® Pentium and Intel® Xeon D-1577 with 16 cores. The highest performing module is based on the AMD EPYC Embedded 3000 series offering embedded edge server designs up to 16 cores within a power envelope of 100 watts. This is a lot of computing power for rugged environments. But edge applications need even more performance, and the PICMG addresses this demand with the new open standard COM-HPC. COM-HPC Server modules will bring such installations a quantum leap in performance. COM-HPC Server modules feature up to 6465 PCIe Gen 4.0 and 5.0 lanes, delivering an 8 times higher overall bandwidth. Also, networking capacity is significantly increased by factor 4 to 8x 25 Gb Ethernet. Moreover, the new Server-on-Modules can have a maximum power consumption of up to 300 watts and can therefore use all the powerful Intel® Xeon and AMD EPYC server processors designed for single processor boards. The standard can even host FPGAS or GPGPUs on the same form factor for massive parallel processing in applications such as radio technology, smart vision and situational awareness or deep learning based artificial intelligence.
What are the emerging applications for HPEC platform?
Much of the demand for robust edge servers resides directly in and at the network and communication infrastructures. Examples are outdoor small cells on 5G radio masts for tactile broadband Internet, or carrier-grade edge server designs. In the industrial field, HPEC platforms power redundant fog servers. In the harsh industrial environment, they are required for Industry 4.0 communication and/or to implement new business models and comprehensive predictive maintenance using big data analytics at the edge. Smart energy networks have similar requirements to reconcile energy generation and consumption in microgrids and virtual energy networks in real-time. In such distributed applications real-time 5G will play a role as well since the required high bandwidths along the infrastructure paths can be provisioned comparatively easily. Autonomous vehicles will benefit too, as they need server performance in the cockpit for processing high bandwidths onboard to analyze data from hundreds of sensors in fractions of a second. Medical backend systems, industrial inspection systems, and video surveillance solutions at the edge for increased public safety are further relevant target markets. Many of these applications will integrate vision. That’s why we engage intensively in this sector to make embedded vision technologies part of our embedded computing platforms.
What are the challenges faced by embedded designers today?
HPEC designs must provide many advanced high-speed interfaces like PCIe Gen 4.0, Gbit Ethernet and USB 3.2 gen 2 x2, to name just a few. At the same time, the number of data lanes increases. With higher frequency signals and more signal lanes, embedded designs are becoming exponentially more complex. Moreover, the routing of the board demands highest expertise with ever increasing frequencies to stay within the limits for a compliant and therefore reliable design. Application-ready super components like Computer-on-Modules and Server-on-Modules can help to reduce the pressure on developers, as they make CPU, memory and high-speed interfaces available as a commercial off-the-shelf (COTS) platform. But engineers utilizing these modules also need support for their dedicated carrier board designs to ensure design compliance. We have addressed this need by offering the same compliance tests for COM-HPC based carrier board/module combinations as we use in house for our own boards and modules. This saves customers from having to invest in their own or external test labs, and allows them to benefit from the expert knowledge of the global market leader in Computer-on-Modules. It is also likely to make it easy to find an expert who can help solve problems in case of carrier board incompliance.
How does this platform enable advanced AI and deep learning applications?
Customers often need to procure technologies from various vendors. This is time consuming and error prone. congatec has built a comprehensive ecosystem for AI and deep learning around its embedded and edge computing platforms to simplify their use in customer applications. This starts with ready-made development kits for embedded vision including pre-validated camera technology from Basler to run embedded vision edge applications with AI-like automated checkout systems in retail, surveillance and access control systems in facility management, industrial inspection systems for quality assurance, augmented reality for maintenance, and image processing in mobile and portable medical technology. Going up on the integration level, customers can benefit from complete platforms for AI solutions. One example is the application-ready AI platform for sparse modeling that congatec offers together with its partner Hacarus. This AI solution needs only a few pictures to learn. Customers benefit from a small and dramatically reduced power footprint compared to conventional deep learning technologies. As a result, OEMs can not only execute but also train their AI at the edge.
What are the advantages of your products in terms of performance and reach?
Our portfolio starts at low power embedded computing technology consuming only a few watts, with boards and modules equipped with ARM application processors featuring a fully integrated OS. From here, our portfolio extends up to the high-end edge server class, covering everything that lies in between those two pillars. Customers can scale up and down as required and always get exactly what they need at congatec. Price and performance are optimized for the specific application, and this on the basis of proven and advanced open standards. Thanks to the standardization, customers benefit from high scalability through simple module changes and best long term availability including second source options. Our reach is global, both from a geographic point of view and in terms of markets as we serve all industries that can benefit from our boards and modules. The most important markets for congatec are edge communication and networking, industrial and robotics, medical and healthcare, test and measurement, as well as vision, kiosk, retail, and infotainment. Smart grids, infrastructures, smart cities, and transportation including autonomous vehicles are other key sectors. For India, we see booming markets especially in industrial automation, medical, energy and education.
What are the new areas where you are focusing on?
As already mentioned, we see a growing demand for extensive vision capabilities combined with AI in many markets. We are investing a lot in this area to offer customers the best in class platform setups. Another topic of utter importance is the need for workload consolidation via virtualization technologies. Here, we have invested heavily in providing full support for real-time hypervisor technology. Customer benefits from virtualization include lower hardware costs and better utilization of the consolidated performance. Addressing both these demands together is pretty hot stuff for the industry and something that is now available with our brand new application-ready, Intel® IoT RFP (Ready For Production) qualified workload consolidation kit for vision based situational awareness. Designed in cooperation with Intel® and Real-Time Systems, the congatec kit targets the next generation of vision based collaborative robotics, automation controls and autonomous vehicles that have to tackle multiple tasks in parallel, including situational awareness utilizing deep learning based AI algorithms, which includes Real-Time Systems’ hypervisor technology as well as Intel® OpenVino software for situational awareness. Both technologies are applicable from high-end to low-end edge computing based on x86 technologies from AMD and Intel. In the low-power sector we offer barrier breaking SMARC 2.1 Computer-on-Modules. Engineers can scale from Intel Atom processor technology down to smallest ARM i.MX8 technology. Overall, we currently offer 24 different modules so that engineers can address their exact price/performance needs. For fastest time to market we have designed carrier boards in the common 3.5-inch form factor that are perfectly suited for many low power applications on the basis of this broad ecosystem.
Can you please discuss your market growth and business plans?
congatec is a high-performance organization and a global leader in embedded computing technologies. We want to continue to grow faster than the entire market and invest a lot in flexibility, customer focus, and our ability to work highly effectively in teams that even integrate customers to simplify the use of our technologies. We expect that a major growth portion will come out of the edge server business as this market is forecast to grow fast and has a high volume. As our offering also includes technologies such as real-time hypervisor support, we see ourselves in a good position to offer unique values to our customers that are highly relevant for the industrial market as well as all the IIoT connected devices. For India, we also see very interesting application areas for our low power i.MX 8 and Intel® Atom platforms. They are great for mobile health equipment, which is not only important in times of COVID-19, as well as for the fastest growing edtech markets where e-learning based products and services are highly demanded, especially these days in times of social distancing. Next there is the power sector, which needs to deploy smart grid technologies in an efficient, cost effective, innovative, and scalable manner. This demand can be perfectly solved with our modular approach, too.
COVID-19 has disrupted most of the industries. What is the impact you are observing and how are you facing this crisis?
Many companies have already implemented supply chain risk management, as the number of disruptions in supply chains has increased steadily and these disruptions have a significant impact on the success of the company. At congatec, we have already had plans for various disasters such as tsunamis and earthquakes in place prior to COVID-19. Our preparations are mainly connected with buffer stocks and second sources, which ensure a sufficient supply over a predetermined period. Moreover, as a fabless company, we are in the comfortable position that our production is more flexible in the event of a crisis. We have been building up the necessary know-how for many years. So today we can leverage an excellent network, which ultimately makes us less vulnerable than embedded computing suppliers whose supply availability depends primarily on large manufacturing centers in Asia. For OEMs, we have an important advice in these insecure times: Leveraging standardized Computer-on-Modules is also an effective means to minimize risk for their own solutions. That’s because Computer-on-Modules such as those offered by congatec are already quite well hardened against the failure of individual suppliers. And since standardized Computer-on-Modules are easily exchangeable, OEMs can also leverage a multi-vendor strategy to reduce the risks on their side even further.