Taking IIoT High Performance Computing to the Edge

[vc_row][vc_column][vc_column_text css=”.vc_custom_1521465181212{margin-bottom: 0px !important;}”]

14 March 2018, Kontron – Edge, fog and cloud applications demand strongly increasing computing power. Traditional server approaches cannot provide the required robustness for operation in harsh industrial environments. Modular systems based on the well-established and tested COM Express modules provide an efficient solution. Type 7 is a new variant, which first time ever satisfies the demands for embedded high performance computing (eHPC).

With centralised cloud computing models unable to keep pace with the growing Big Data volumes and real-time processing requirements of the IIoT, high performance edge and fog applications are being increasingly deployed in the industrial sector. These demand a new category of robust embedded high performance computing (eHPC) servers offering exceptional processing power and the ability to control, compute and connect multiple applications on one platform.[/vc_column_text][mk_padding_divider size=”8″][vc_column_text disable_pattern=”false” css=”.vc_custom_1521465948452{margin-bottom: 0px !important;}”]EXECUTIVE SUMMARY
Many organisations in the industrial sector have been aware of the concept of smart connected machines for a number of years. However, in contrast to most business and IT applications many production tasks cannot be easily outsourced to the cloud, even though it would be preferable for reasons of simplification, cost savings, and ease of maintenance.

This is in spite of new infrastructures such as time sensitive networks (TSN) because of the physical distances and associated latencies involved: making real-time control out of the cloud impractical. Furthermore, for reasons of security, data privacy and compliance, many companies are uncomfortable with their entire production data and intellectual property being stored and processed off the premises.

As a result those organisations wishing to connect plant, machinery and other equipment involved in the manufacturing and production process have resorted to
building complex and expensive proprietary systems. Often, however, these cannot be monitored and controlled from a common platform due to many mechanical legacy devices being without digital capability. There is also the sheer number of other incompatible devices, controllers and control languages to consider. For many, therefore, it has been challenging to leverage the IIoT for streamlining operations and realising the many business benefits available.

However, things are changing with the rapidly falling cost of increasingly sophisticated sensors and analytical software; and crucially the availability of powerful edge and fog computing systems. Here servers reside on the edge of the IIoT for enabling much more efficient capture, processing and real-time analysis of operational data – at the point of activity – before transmission as necessary to a centralised cloud for further analysis and archiving.

These developments have created the ‘Perfect Storm’ conditions for the proliferation of the IIoT, not only by reducing the cost of acquiring data, but also in creating much more flexible and responsive localised high performance computing environments. They are becoming essential for processing Big Data volumes in real-time.

With the growing demand for highly connected edgebased environments this paper discusses why more powerful embedded high performance computer (eHPC) technology is becoming necessary; for managing and optimising availability, performance, reliability, scalability, and security. This has subsequently led to a new COM Express standard revision (3.0) which has been specifically designed to accommodate new server-type processor technology. As part of this discussion, an introduction to Kontron’s new range of powerful, fit for purpose eHPC COM Express Type 7 solutions for edge computing is also included.

INTRODUCTION: eHPC AT THE EDGE
The Industrial Internet of Things (IIoT) offers many industrial organisations the opportunity of realising significant business benefits. By enabling production, assembly and research operations to be carried out more efficiently, reliably and safely they can achieve greater cost savings, higher customer satisfaction and sharper competitive edge.

But this requires controlling, processing and analysing exponential data flows in real-time to meet the growing requirement for actionable information; to streamline production, for example, or enable pre-emptive maintenance on plant and equipment.

However, total reliance on a centralised cloud computing model is largely unsuited to helping organisations realise such IIoT derived benefits. This is down to network latency issues impacting on real-time data collection and analysis due to the often considerable distances between cloud data centre locations and local devices and sensors. There are also concerns over sending sensitive data across various networks into the cloud.

Therefore, to ensure the IIoT can make data acquisition fully accessible to industrial organisations engaged in a wide variety of activities, it has become necessary to find a more efficient and faster way of controlling, processing and storing the large volume of data being generated.

In response to these technological challenges, decentralised edge eHPC environments are increasingly being deployed. These are located close to the smart devices and local facilities that act as data collection points, be they factories, drilling platforms, research centres, farms, office blocks and so on. In turn this facilitates the operation of compute, storage and networking services between the end devices and centralised cloud data centres.

By performing much of the data processing, control and management of local applications in close proximity to the devices or sensors, latency is greatly reduced and application responsiveness optimised, enabling real time processing of mission critical data. At the same time, only less time sensitive data is transmitted to the cloud for further processing and analysis which reduces overall network congestion.

For these reasons, a combination of edge and cloud computing is quickly becoming the accepted architecture for working with IIoT data. This model is also referred to as fog computing: extending the cloud down to the local level to combine centralised and distributed computing resources into a single architecture; allowing edge devices to communicate with one another and with the cloud.

EDGE-BASED eHPC: NEW IMPLICATIONS FOR EMBEDDED SYSTEMS
The accelerating demand for more sophisticated high performance computing at the edge, combined with smarter IIoT devices, is creating a number of embedded computing challenges. One of the most important is how to meet the growing requirement for more scalable processing power and network bandwidth; to keep pace with the rapidly growing volume of data being captured from more and more devices and equipment.

In addition, there are the inherent physical space constraints to consider, bearing in mind embedded systems are increasingly integrated into the machines and devices themselves. Take for example, robot controllers, or test and measurement systems.[/vc_column_text][/vc_column][/vc_row][vc_row][vc_column][mk_content_box heading=”Design Requirements” icon=”mk-icon-angle-right”][vc_column_text css=”.vc_custom_1521466419868{margin-bottom: 0px !important;}”]

  • Data analysis and real-time processing on a single platform
  • Small footprint design
  • High scalability
  • Faster networking capabilities
  • Powerful multicore processing
  • Lower power consumption
  • Reliability, Availability, Manageability
  • Integration of all interfaces at single board level
  • Robust design for harsh environments
  • Provide security of IIoT edge software applications
  • Long term availability – secured upgrade path
[/vc_column_text][/mk_content_box][/vc_column][/vc_row][vc_row][vc_column][vc_column_text css=”.vc_custom_1521466690836{margin-bottom: 0px !important;}”]RISING TO THE CHALLENGE
Standardised computer modules (COMs) such as COM Express are well-proven as an essential building block in many embedded industrial computing applications. For example, the COM Express Standard Type 6 more than adequately addresses many of the embedded computing needs of systems designers and users. This has been helped by a some embedded solutions manufacturers offering essential applications pre-integrated on carrier boards as well as custom interfaces so that all necessary functionality is easily available at single board level. Furthermore, COM Express Type 6 has hitherto offered the widest scalability with its compact “Basic” size form factor of only 95 x 125 mm² able to use processors with a TDP (thermal design power) of 50 watts or more without any problems.

However, to ensure processing power and network performance is optimised for controlling future edgebased eHPC applications, the PCI Industrial Computer Manufacturers Group (PICMG) has recently redefined its highly popular COM Express standard. The latest revision (3.0) has resulted in the introduction of COM Express Type 7 which allows the design of highly scalable servertype modules offering multicore parallel processing and excellent networking capabilities on one platform.

COM EXPRESS TYPE 7
Derived from and complementary to COM Express Type 6, the new Type 7 standard is ideally suited to maximising the multicore processing power of the latest Intel® Xeon® D-1500 and Atom® C3000 processor family of system on chip (SoC) processors.[/vc_column_text][/vc_column][/vc_row][vc_row][vc_column][mk_content_box heading=”COME TYPE 7 – KEY REVISIONS ” icon=”mk-icon-angle-right”][vc_column_text css=”.vc_custom_1521466724409{margin-bottom: 0px !important;}”]

  • Four 10GBit interfaces
  • 32 PCI Express® lanes
  • A complete set of NC-SI sideband signals
[/vc_column_text][/mk_content_box][/vc_column][/vc_row][vc_row][vc_column][vc_column_text css=”.vc_custom_1521467224679{margin-bottom: 0px !important;}”]BENEFITS
The scalability provided by the modular design approach of COM Express Type 7 makes it ideally suited for meeting the increasingly demanding processing requirements for eHPC applications at the edge. The flexibility of multicore empowers system designers to assign multiple tasks on a single board computer (SBC) using dedicated processors for each one, therefore significantly increasing performance.

Equally important, the availability of four 10GBit network interfaces offers considerably more scope for accommodating high speed Ethernet connections. An additional eight PCI Express lanes solves the increasing demand for connecting I/O devices with high data throughput. Other key features include NC-SI sideband signals providing the ability to perform remote diagnostics and pre-emptive maintenance activities in combination with a Board Management Controller (BMC) on the baseboard. In order to accommodate these additions and changes, the Type 6 pinout was partially modified to no longer support any audio and graphics interfaces since these are largely expendable in edge-based IIoT environments. There are also only four, rather than eight USB 2.0 ports;; and just two of the four SATA ports remain.

KONTRON: AT THE LEADING EDGE OF IIoT- READY eHPC SOLUTIONS
As one of the world’s leading embedded computer technology (ECT) providers, Kontron is renowned for the reliability and quality of its IoT-ready controller, module, gateway and system products. The company provides extensive support along with notably long-term product availability, allowing industrial customers to benefit from a high level of investment security.

Kontron is a longstanding member of the PICMG and for over 15 years has been instrumental in the ongoing evolution of COMs, including the creation of the latest COM Express Type 7 standard. Through the design of standardised COM Express® modules, the platforms are easily exchangeable offering the most flexibility for customers designing them into different size embedded devices. Moreover it provides a secure approach allowing easy upgrade possibilities with future upcoming compatible module designs. As a true partner, Kontron delivers all of the needed support to help customers reduce time-to-market and gain a competitive advantage.

KONTRON COME TYPE 7 eHPC PORTFOLIO

Kontron’s new powerful COM Express Type 7 eHPC module portfolio (COME-bBD7 and COMe-bDV7) provides a broad feature variety and high scalability for various application requirements:

  • 2 to 16 CPU cores with Intel® Atom® C3000 or Intel® Xeon® D-1500 processor family
  • Max 45 watts TDP
  • Up to 4x 10Base-KR interfaces
  • Up to 64 GByte DDR4 memory
  • Extended temperature range variants of -40 °C to 85 °C
[/vc_column_text][vc_single_image image=”6433″ img_size=”full” alignment=”center”][vc_column_text css=”.vc_custom_1521467639808{margin-bottom: 0px !important;}”]For maximum network data performance, the COMebBD7 supports 2x 10GBaseKR Ethernet interfaces, while the COMe-bDV7 provides 4x 10GBaseKR Ethernet interfaces. The Intel I210IT 10/100/1000 Mbit Ethernet controller integrated on the COMe Type 7 modules is ideally suited for network management purposes. Additionally, the modules can be connected to an external BMC (board management controller) present on the baseboard via NC-SI (Network Connect Sideband Interface). The BMC allows monitoring of the COMe Type 7 modules and supports remote management in order to enable, for example, pre-emptive maintenance.

Up to 32 GByte DDR4 ECC memory on the COMe-bBD7 and 64 GByte DDR4 ECC on the COMe-bDV7 ensures Server Class RAS (Reliability, Availability, Serviceability). Beside SPI, LPC, SMB, Fast I²C, other features included as standard on the module are RTC and staged watchdog, 2 serial interfaces, 4x USB 3.0 / 2.0, and 2 SATA3 interfaces with 6 Gb/s.

NEW CARRIER BOARD[/vc_column_text][vc_column_text css=”.vc_custom_1521467929761{margin-bottom: 0px !important;}”]The COMe Evaluation Carrier T7 Board has been newly developed especially for the Kontron COMe Type 7 modules. It comes with an ATX form factor (305 x 244 mm²) and forms a turnkey server development platform together with the module. It supports, for example, 4x 10GbE, 32 PCIe lanes, 4x USB3.0, 2x SATA3, 2x RS232, GPIO, and an optional IPMI 2.0 compatible BMC with KVM support via an adaptor card.

Full design documentation of the evaluation carrier is included for fast baseboard development.[/vc_column_text][/vc_column][/vc_row][vc_row][vc_column][mk_content_box heading=”MAXIMISING IIoT SECURITY” icon=”mk-icon-angle-right”][vc_column_text css=”.vc_custom_1521468002643{margin-bottom: 0px !important;}”]

  • Kontron’s COMe Type 7 modules are the embedded computing industry’s first edge server-class COMs enabling by default a comprehensive on board security solution which encrypts a software application’s source code in a way that protects IP and makes reverse engineering impossible
  • Kontron APPROTECT is a combined hardware and software solution that in addition to the TPM 2.0 (Trusted Platform Module) chip includes an embedded hardware security chip and a software framework, providing full application level protection
  • The new security solution is also available for all Kontron products either integrated in new products starting with the 6th generation Intel® Core™ and Intel® Xeon® processors as well as latest generation Intel Atom®, Intel® Celeron® and Intel® Pentium® processors, or by using an upgrade kit
[/vc_column_text][/mk_content_box][/vc_column][/vc_row][vc_row][vc_column][vc_column_text css=”.vc_custom_1521471360037{margin-bottom: 0px !important;}”]IIoT ESSENTIALS: EDGE AND FOG
Edge and Fog computing are complementary and together effectively manages computation, communication, control and storage along the continuum from cloud to connected devices (things).

On the one hand, Edge computing optimises cloud computing systems by performing data processing at the edge of the network, near the source of the data. This reduces the communications bandwidth needed between sensors and the central data centre by performing analytics in close proximity to the source of the data.

At the same time, Fog provides the flexibility for managing what data needs to be sent to the cloud, and what should be analysed locally at the edge. For example, certain applications will only require simple analysis and data does not need to be sent to the cloud which would only increase latency and costs while also unnecessarily delaying decision making. With the help of powerful eHPC technology the data processing can instead occur at the edge to provide results in real-time.

Furthermore, on the factory floor, Fog bridges the gap between modern Enterprise IT and Operational Technology (OT), enabling scalable computing at the edge, resource virtualisation, modern application management and data interoperability middleware.[/vc_column_text][/vc_column][/vc_row][vc_row][vc_column][mk_content_box heading=”LIVING ON THE EDGE: IIoT APPLICATIONS” icon=”mk-icon-angle-right”][vc_column_text css=”.vc_custom_1521471574794{margin-bottom: 0px !important;}”]Edge-based eHPC is being applied to a growing and diverse range of IIoT applications. Here are a few examples: Edge-Smart Factories:

  • Providing holistic real-time operational views of legacy and modern plant and equipment – ensuring customer satisfaction from maximum reliability and avoidance of lost productivity
  • Enabling predictive maintenance where operators and technicians can more quickly and accurately monitor performance metrics to anticipate when a particular part is going to fail; replacement parts can be delivered before an issue arises; personnel can be alerted immediately to address machine malfunction issues before production is affected
  • Remote real-time connections to machine users allowing ‘on the fly’ discussion and swift resolution to potential issues; analyse machine data live and respond with appropriate solutions
  • Handling software updates remotely rather than on-site to avoid business disruption

Smart Energy Production:

  • Reducing costs by improved power transmission and distribution efficiencies from access to more accurate, actionable information – in real-time
  • Optimizing Renewable Energy’s contribution to the grid – access to real-time information for providing up-to-the-second situational awareness of grid performance; actively manage distributed energy resources (DERs) from renewable providers to streamline power and energy flow
  • Maximising Windfarm productivity by enabling onsite edge software to perform real-time analysis on actual wind conditions; as wind direction alters communicate to each wind turbine, adjusting its position appropriately. With data not having to be sent to the cloud, the processing time to instigate this change is faster, maximising windfarm production
  • Enabling Oil and Gas drilling in areas that were previously considered inaccessible at any cost; from the Arctic cold to the heat of the desert, gaining maximum ‘intelligence’ from thousands of connected sensors measuring seismic information, machinery performance, environmental conditions, oil flow rates and pressures; what used to take decades to examine and understand, is now adding production value in a matter of weeks
[/vc_column_text][/mk_content_box][/vc_column][/vc_row][vc_row][vc_column][vc_column_text css=”.vc_custom_1521471843198{margin-bottom: 0px !important;}”]SUMMARY
Gaining business efficiencies and competitive advantage from IIoT-derived Big Data is the future. With this, more and more data must be reliably recorded and processed physically close to where plant, machines, and other devices are located. As a result edge computing is increasingly being deployed to optimise network and applications responsiveness. It is inherently more suited than centralised cloud infrastructures for rapidly processing and analysing data in real-time.

At the same time, the increasing demand for more powerful and sophisticated edge computing servers for IIoT applications has necessitated new thinking on embedded systems design. In 2017 this led the PICMG to introduce its latest COM Express Type 7 standard. In turn Kontron has been quick to respond with its first range of COMe Type 7 multicore processor embedded high performance computer (eHPC) solutions. Enabling powerful data analysis and real-time processing on one platform these are the first on the market to offer sophisticated on-board security.

For more information on Kontron’s embedded systems please get in touch with us via the contact form on the page below:[/vc_column_text][vc_btn title=”Visit Kontron” color=”success” align=”center” i_icon_fontawesome=”fa fa-chevron-right” button_block=”true” add_icon=”true” link=”url:https%3A%2F%2Fwww.l2tek.co.uk%2Fsuppliers%2Fkontron%2F|||”][/vc_column][/vc_row]