Embedded server technology for the edge of smart grids
Tencent’s energy IoT partner wants to change how distributed energy production and consumption are managed by implementing a new monitoring, management and control layer of distributed edge servers. These new small and rugged but nevertheless tough and powerful high-speed devices can connect in real time to all the decentralised devices relevant for energy management.
First implementations of this new edge sever layer can be found in Shanghai Business Park. They are built on real-time capable server-on-modules from congatec, which offer a wide performance and interface scalability from Intel Xeon to Intel Atom C3xxx processors to meet the given local needs in a flexible and highly automated way.
In the information industry, the generation and exchange of data is one of the most important production factors and value drivers. This is even truer for the energy sector, as there is a massive trend away from central power generation to a distributed and highly fragmented energy grid with various energy consumers and producers. Such a comprehensive network topology raises massive real-time information and synchronisation demands to ensure optimal grid balancing.
On top of this come energy-saving aspects, for example, to reduce the costs of energy-hungry productions, to save the environment and to protect the climate. As a result, even more information is needed to improve energy efficiency across households, factories, cities and even entire nations. Finally, there is a call for better power delivery management, which requires the installation of new services such as remote meter management to reduce costs or to offer volume-based tariffs with special peak load controls using artificial intelligence (AI) to improve user experiences and to more efficiently balance the entire power consumption throughout the day.
With all those various needs, many heterogeneous solutions for nearly every specific demand are deployed all over the world to support energy companies in achieving intelligent operational processes. But until today, there was no holistic approach in sight where all relevant energy management devices are real-time connected by local edge server platforms. Such a new local layer would be capable of managing the entire energy ecosystem of distributed power devices.
It could provide, for example, core applications for multi-energy collaboration, demand-side response, comprehensive environmental management, efficient operations and maintenance management, as well as user-centric energy services for companies such as China’s regional integrated energy service companies (ESCOs). By integrating advanced technologies such as Tencent’s distributed service framework (TSF) and energy IoT as well as big data analytics and visualisation, it would help ESCOs to optimise operational and maintenance efficiency and save energy.
To offer such a comprehensive service, the Chinese energy market experts from Tencent’s energy IoT partner have combined the technical capabilities of the Tencent Cloud with large objects, and are collaborating with the power grid, industry-leading service providers, universities and research institutes to provide the integrated ESCOs with a service- and feature-rich edge server infrastructure. The solution is designed to support the rapid transformation of traditional energy services companies by innovating and integrating the next generation of information and energy technology. Once the service is entirely available, it will support thousands of power devices, hundreds of power protocols and online editing of common standard protocols as well as dynamic completion of common standard device types.
All this cannot be done by central clouds or central servers, as the limiting factor is a required response time below 10 ms so that the control system can accurately react to the varying power sources and sinks. In a central cloud topology, data travel times are usually around 150 to 200 ms from the source to the central cloud and back. Locally deployed edge servers can reduce data exchange times between power equipment and controls to 2–5 ms. This significantly shortens the latency and reaction times, thereby optimising the performance of smart grid applications.
That’s why Tencent’s energy IoT partner pushed the process to bring the cloud to the edges of the Chinese energy grid. There was a need to build an edge server system that is capable of bridging all the heterogeneous local protocols and interconnects; delivering high computing power for protocol transcoding, AI and real-time control of all the many different devices in the field; and offering the required bandwidth for billions of messages from and to the devices. Additionally, it has to be designed to withstand the harsh environmental conditions in the field, for example, offering resistance against high and low temperatures as well as shocks and vibrations, plus it should offer server-like remote management features for reliable operation and reduced maintenance efforts. As those systems also serve as a local IoT gateway/hub for exchange with the central Tencent Cloud and additionally need to enable fast deployment of energy devices via QR or barcode scans on mobile devices, they must also provide high bandwidth connectivity to the central cloud.
Today, a one-click scan code helps system integrators to support fast power device deployments. With such a scan, they can quickly bind and parameterise the power device to the edge server and its cloud (aka fog). In addition, non-contact sensors are used for continuous installation and debugging in deployment mode. A virtual device library assists the user in device selection and evaluation. With all these valuable functions and computer-assisted workflows, the device installation and deployment cycle has been reduced from one day to one hour. The upper layer of the installation uses certificates or key authentication methods to ensure secure access to devices.
Since May 2018, the first field tests of the new energy edge server platform have been finalised and first real-world applications are now going to be deployed in Shanghai Business Park where the real-time collection of all types of data is executed for dynamic monitoring, energy analysis, cost accounting, performance evaluation and reporting functions such as publication, energy management optimisation and energy-saving promotions. The scope of services includes technical load and security issues as well the management of purchasing and sales contracts, including risk and energy efficiency management.
The future goals are even more impressive and driven by the massive, intelligent and fast interconnections: with the combined computing power of Tencent Cloud’s global data centre and the powerful edge, such a distributed edge server layer can support billions of access points to energy equipment as well as the reliable transmission of trillions of messages and decisions. All this can be realised in real time thanks to the decentralised edge server logic to maximise the energy efficiency from households and microgrids to smart cities and virtual grids at lowest costs. It is estimated that approximately 1.26 million edge severs are required to build a nationwide edge server infrastructure for China, calculating that per 1000 inhabitants one single edge server is needed. So there are quite a few edge servers to deploy over the next years until that goal is reached, but at least the first cornerstone has already been laid with the recent installation.
For the adaptive hardware of the new class of edge devices, the edge server designers have chosen a flexible system design based on COM Express Type 7 modules. These modules offer server-grade performance and functionality on a standardised form factor plus cooling solutions. The modules come as application-ready super components, including core function units like CPU, control hub, memory as well as USB, PCIe and Ethernet connectivity. COM Express Type 7 modules are the first modules to specify multiple 10 Gigabit Ethernet connects in this context. For fast connectivity they offer an increased count of up to 32 PCIe lanes to add, for example, fast NVMe storage as well as GPGPUs for AI based on deep neural networks. The application-specific design is executed via a carrier board that holds the application-specific connectors and peripherals. This makes it possible to leverage one board design across different performance classes by just exchanging the module. This strategy also brings benefits for future upgrades, as more performance can easily be integrated by simply swapping the module against a more powerful variant.
First systems are based on modules with Intel Xeon D15xx processors with up to 16 cores and 32 threads. These processors offer a high per core performance thanks to Intel AVX2 extensions increasing performance when the same operations are performed on multiple data objects — relevant for AI for example — as well as Intel TSX-NI, which is focused on multi-threaded performance scaling and helps to make parallel operations more efficient. All these features result in an impressive performance per watt with a high throughput and overall server-grade performance. Alternative configurations are based on the Intel Atom C3xxx processors. Featuring up to 16 cores, those processors are the ideal fit for all installations that have to handle smaller package sizes in parallel while consuming as little energy as 11 W TDP.
Both modules offer ECC RAM for high data integrity and support the extended temperature range from -40 to +85°C for good reliability even under extreme conditions. The hardware vendor is congatec, who — in cooperation with its Chinese VAR as well as Intel — is providing the entire hardware platform as a boxed system including all required tests and services.
“It is great to work with congatec on this energy edge server project which will enable brand-new services,” said Zhang Zhidong, CTO of Tencent’s energy IoT partner. “As they provide us with a local contact here in China, it is very easy and comfortable for us to synchronise our needs. And with congatec working at the forefront of the COM Express Type 7 standardisation, we couldn’t have chosen a better supplier.”
A cortical vision device could one day help restore sight to the blind, thanks to the use of...
Edge computing used in conjunction with cloud computing is becoming an increasingly indispensable...
Cutting-edge innovation, combined with agility and digitalisation, are likely to be key factors...