Edge Computing Unleashes Physical Potential

Two Themes of Hannover Messe 2018

In April, 2018, Germany held the Hannover Messe Industry Exhibition, an annual event that identifies industry trends. At Hannover Messe 2018, two Operational Technology-Information and Communications Technology (OT-ICT) convergence topics attracted the most attention:

• Industry integration: Connection and collaboration. Why is industry integration the main theme for today’s manufacturing environment? Customers demand integrated solutions that require vendors to collaborate with each other; however, there remain a number of closed coalitions in the manufacturing industry that are not willing to collaborate with each other. Without data exchange, companies within these coalitions cannot extract more value from their service data. Today, open standards, such as the OPC Unified Architecture (OPC-UA) and Time-Sensitive Networking (TSN), are undergoing rapid development. Horizontally decoupled open platforms are widely applied and promoted in the manufacturing industry. Vendors can connect to each other to perform cross-field collaboration. Open connections and collaboration enable players along the industry chain to flexibly select partners, focus on their core business, achieve a win-win ecosystem, and create more space for innovation. At Hannover Messe 2018, Huawei actively responded to the event’s theme and the industry players’ common requests, designed the TSN network and OPC UA test bed for six major application scenarios, and released the network and test bed with more than 20 international organizations and highly regarded vendors. These organizations and vendors included, but were not limited to: Alliance of Industrial Internet (AII), Avnu Alliance, Edge Computing Consortium (ECC), Fraunhofer FOKUS, National Instruments (NI), Schneider Electric, and B&R Industrial Automation.

• Ubiquitous ‘digital twin’: The physical and digital worlds. The exhibition booths for many mainstream vendors focused on ‘digital twin’ technology. Based on connections, cognition, and perception to the physical world, we can create a mirror in the digital world to enable bi-directional data and information flow. Conversely, we can leverage the innovative technology we create in the digital world to better understand and optimize our physical world. Over the last decade, network, computing, and storage — the three pillars of the ICT industry — experienced exponential growth in technical and economic feasibility. Network bandwidth increased by 1,000 times, while cost per bit decreased by 97.5 percent. The cost of computing chips fell by 98.4 percent. The capacity of a single storage disk increased 10,000 fold, while its cost decreased by 94.2 percent. Next-generation IT technologies such as Big Data and AI experienced leapfrog development and were extensively implemented, ushering in an era in which digital ICT unleashes the potential of the physical world.

Collaboration between Edge Computing and Cloud Computing

How do we connect the digital and physical worlds? Directly connecting the two worlds pose various challenges, such as high latency, high bandwidth costs, low reliability, and security and privacy issues. Many of these challenges can be addressed by migrating intelligent applications from the cloud to the network edge. We need to connect the two worlds using edge computing in a location close to the physical world. Edge computing is a distributed open platform that converges network, computing, storage, and application abilities. The platform offers local edge intelligent services to meet digital industry transformation’s requirements for agile connection, real-time service processing, data optimization, intelligent applications, and security and privacy protection. Cloud computing focuses on centrally processing and storing a massive amount of data. Real-time response is not cloud computing’s core objective. Edge computing uses distributed computing to enable the live, automatic, and collaborative operation of physical equipment.

Collaboration between edge computing and cloud computing can improve digital industry transformation in the following aspects:

• The physical and digital worlds, once separate from each other, are now converged for more productive collaboration.

• Enterprises make business decisions based on digital models, instead of managers’ personal experience.

• All workflows are connected together, allowing data-based, end-to-end collaboration.

• Enterprises no longer toil alone; they work with ecosystem partners to achieve innovation.

Recognizing the important role that edge computing plays in digital industry transformation, Huawei started to release edge computing products and services in 2016, including Edge Computing Internet of Things (EC-IoT) gateways and open cloud-edge collaboration services.

On edge computing nodes like network edges, edge gateways, and edge servers, the local edge computing platform creates and runs the models and applications developed on the cloud to serve physical objects such as engineering machines, production lines, and transportation vehicles. The edge gateway is responsible for bus protocol conversion, and data convergence and optimization. The edge server analyzes and stores mass volumes of service data in real time. Edge computing nodes can work together based on regions and services to build an edge cloud that supports elastic horizontal expansion. The edge cloud inherits the traditional advantages of public and private clouds regarding the capability to scale easily. In addition, it provides a lightweight, intelligent, distributed architecture that features easy O&M to enable flexible deployment, intelligent distribution, and application networking. In public and private cloud data centers, the edge cloud also gives partners and end users centralized cloud computing services, such as AI, application DevOps, and Big Data services. With edge clouds in place, enterprises can centrally develop and deploy applications and models.

Huawei’s open cloud-edge collaboration services have been applied in smart manufacturing, smart water, smart building, and many other industries. In smart manufacturing, robots are used to ensure product quality. Traditionally, enterprises would assign quality inspectors to manually check their products, which is inefficient and prone to false detection and missed faults. Edge-cloud collaboration and deep learning allow robots to perform the inspection, improving product yields and detection efficiency. In smart water scenarios, pipe pressure sensors communicate with each other to simultaneously detect pressure changes in multiple pipes. When pressure exceptions are detected, the system data exchanged between modules is used to ensure timely, accurate fault detection. Additionally, the system can also detect pipeline risks and automatically control valves to minimize risk and prevent losses.

Three Steps to Promote Industrial Implementation of Edge Computing

Huawei is also committed to promoting the healthy development and mutually beneficial cooperation of the edge computing industry. In 2016, Huawei established the Edge Computing Consortium (ECC) with the Shenyang Automation Institute of the Chinese Academy of Sciences, the China Information Communication Research Institute, Intel Corporation, ARM, and iSoftStone Information Technology (Group). Huawei is now cooperating extensively with a growing number of important industry organizations, including the Industrial Internet Consortium to jointly promote digital industry transformation.

The ECC uses a three-step strategy to continuously iterate, optimize, and promote the industrial implementation of edge computing:

• Step 1: Define the edge-computing reference architecture. The open edge-computing platform needs seamless collaboration along the industry chain to promote ideas and commercial implementation. The reference architecture will provide a unified top-level design to help the industry reach consensus.

• Step 2: Build and release an open edge-computing platform. Under the guidance of the reference architecture, enterprises along the industry chain cooperate with each other to build an open edge-computing platform, provide pre-integrated software and hardware system modules, and open interfaces to alliance members, driving the commercial use of their products and solutions.

• Step 3: Define the reference architecture of vertical industries based on best practices and build industry-oriented ecosystems. Promote large-scale implementation in industries based on test bed demonstration and test certification.

Based on its industry practices, ECC released the model-driven edge-computing reference architecture 2.0 in 2017.

Based on these models, we can digitally monitor the physical and digital worlds to:

• Enable collaboration between the two worlds.

• Facilitate cross-industry collaboration.

• Reduce system heterogeneity and simplify cross-platform migration.

• Give effective lifecycle support to service systems.

The architecture has the following features:

• Intelligent services based on the unified, model-driven service framework. The software development interfaces are consistent, and deployment and operations are automatic.

• Service orchestration defines Enterprise-to-Enterprise (E2E) service flows, enabling services to drive the physical world.

• The edge cloud allows automatic visual deployment and operations of the physical world and elastic expansion of edge computing resources.

• Edge computing nodes are compatible with various heterogeneous connections, support real-time processing and response, and provide services such as integrated software and hardware security and intelligent data analysis.

The edge computing reference architecture offers model-based open interfaces at all layers. The architecture provides intelligent services to cover the entire service process using vertical management, data lifecycle, and security services.

To sum up, Huawei will provide partners with next-generation ICT, such as edge computing and cloud computing, and ‘dance’ with ECC partners to write a symphony of digital industry transformation.