How does Intel layout the industrial Internet

2022-08-15
  • Detail

How Intel arranges industrial interconnection

with the gradual maturity of 5g, edge computing, artificial intelligence and other supporting technologies, the boundary between interconnection and industry is gradually blurred, and the industrial interconnection of the two is no longer far away, which has attracted the attention of all parties in the context of manufacturing industry upgrading. Intel, the chip giant, also saw the opportunity and actively deployed relevant technologies and industrial applications

manufacturing is one of the key scenarios of Intel's layout in the IOT market. In this manufacturing market with a long chain and complex scenarios, what does Intel use to knock

leverage the industrial Internet market with the ecological brand

pc is a general standardized market, as well as the server market. As long as a chip is made, it can basically take all the world. When there is no rise of tablet, Intel is all over the world in the it standardized product market. Everyone knows the story of Intel losing the market in the tangle of the rise of mobile Internet. So when the IOT market has not fully developed, Intel will not hesitate to make a high-profile layout to avoid making market mistakes again. However, IOT is a fragmented market. How Intel, which pursues large volume, wide range, standardization, scale and market maximization, can adapt has become a headache

however, Intel seems to have great determination. In 2013, it changed the name of the embedded business division to the IOT business division, and took industry as one of the four key scenarios to test the water, because this market has huge demand and purchasing power

how to start the fragmented market of industry and how to find the large-scale demand are issues that Intel urgently needs to consider. Christineboles, vice president of Intel IOT business unit and general manager of industrial solutions business unit, gave Intel's path is to create reference designs and provide reference platforms so that customers and partners can see how to use Intel's technologies and solutions to improve quality, efficiency, transformation and upgrading in specific scenarios

in the words of Chen Wei, general manager of Intel IOT business unit in China, providing reference design is from the perspective of solutions, so that a chip can be quickly applied in vertical industries. He further explained that providing reference design means that Intel designs some solutions for solution manufacturers' reference, which plays an enlightening role, so that chips can quickly enter the application of vertical industries. Providing a reference platform means that Intel plans to define some complete industrial interconnection components, including sensors, switches and back ends, so that different partners can see a complete closed loop, and then decide their own and other partners' roles in the industrial interconnection, and combine them to form the whole solution, so as to provide market ready solutions for end users

leveraging industrial interconnection, Intel is still playing the ecological card. Although the ecology of industrial interconnection is very different from that of PC and server, for example, the ecological chain of industrial interconnection is much longer than that of PC market, the industrial interconnection ecology of each subdivision field can be quickly combined and connected with reference design and reference platform. If partners such as small and medium-sized bridges, three-dimensional parking garages, storage facilities, etc. see a complete business closed loop and a complete business model, they will actively participate in it, and this ecosystem will continue to be enriched. In the second year of the establishment of Intel IOT business unit, the turnover showed rapid growth, but the profit fell, which many people were puzzled. It seems easy to understand now, because to pry a market, we need to cultivate an ecosystem, which requires time and cost, and a lot of investment

general platform to solve the problem of fragmentation

for the problem of fragmentation in the field of industrial interconnection, Intel's solution is to try to take into account the replicability of products and solutions to benefit more users in different industries. Zhang Yu said that on the one hand, Intel actively participates in the cooperation between standards organizations and the open source community to jointly build open source solutions for common problems, so that they can be applied and integrated into the solutions by a wider audience; On the other hand, it provides complete related products from the bottom chip to the upper software to help industry users solve practical problems

it is reported that in terms of technology application, Intel combines its chip advantages to launch solutions in chips, boards and software tools, including visual acceleration chips, software development packages that support accelerating high-performance computer vision applications and deep learning reasoning, etc. in terms of software, Intel has also been tracking cutting-edge AI technologies such as network compression, and can also provide reference cases for use scenarios in vertical fields, Help developers build applications quickly

taking the visual field as an example, network hard disk video recorders and video servers have higher requirements for computing power, and can accept higher power consumption. In response to such needs, Intel has released a series of vision accelerator products based on intelmovidius myriad x vision processor and arria10fpga. The movidius scheme can integrate 8 to 16 movidiusmyriad x chips on a board, providing 8 to 16t computing power. Users can choose different configurations according to the performance indicators of their edge devices

at the same time, in order to help developers develop machine vision and deep learning applications, Intel also released the openvino toolkit, which supports accelerating high-performance computer vision applications and deep learning reasoning, helps developers and data scientists accelerate computer vision workloads, simplifies deep learning deployment, and easily realizes edge to cloud heterogeneous execution in various Intel platforms

the development of industrial interconnection is inseparable from the construction of a complete ecosystem. At present, Intel has worked closely with alliance organizations such as the global industrial IOT alliance, the open Internet foundation, the open fog Computing Alliance, the edge computing industry alliance, and the industrial Internet industry alliance to integrate upstream and downstream partners of the industrial chain into one platform to jointly explore solutions

hit the pain point of the industry and launch a full stack solution

for the industrial trend of edge cloud collaboration, Intel has launched a full stack solution for edge computing and AI products covering chips, boards and software tools with its leading edge computing and AI technology. It can be used in various devices from smart cameras, smart network video storage to smart video servers

in 2018, Intel released a new generation of visual acceleration chip Intel movidius myriadx. This chip with an area of only 8.8x8mm and power consumption of only 2W can provide 1t computing power, so as to confirm whether the detected sample gap is qualified. Now it accelerates the convolution layer, full connection layer and activation function in the convolution neural network. In practical applications, myriadx can well meet the requirements of power consumption. At the same time, its computing power is also enough to meet the needs of processing all the way HD video

compared with low-power chips, network video recorder (NVR) and video server have higher requirements for computing power, and the acceptable power consumption is also higher. For the needs of elastomers and adhesives, Intel released the Intel vision accelerator product family based on Intel movidius myriad x vision technology processor and high-performance Intel arria 10fpga. Take movidius' solution as an example. This solution can integrate 8 to 16 Intel movidiusmyriad x chips on a board, providing 8 to 16t computing power. Users can choose different configurations according to the performance indicators of their edge devices

in order to help developers develop machine vision and deep learning applications, Intel also released the openvino toolkit. Openvino toolkit supports accelerating high-performance computer vision applications and deep learning reasoning, helps developers and data scientists accelerate computer vision workloads, simplifies deep learning deployment, and easily realizes edge to cloud heterogeneous execution in various Intel platforms

Dr. Zhang Yu, chief technology officer and chief engineer of Intel's IOT business unit in China, said that from the actual production situation, Intel's long-term industrial level in China is uneven, mostly in the stages of 1.0, 2.0 and 3.0, and there is little high toughness in the strict sense; PETG is highly transparent 4.0. There are many reasons, such as insufficient data collection capacity, lack of interfaces, and difficulties in interworking between different devices

however, Zhang Yu believes that the industrial industry inherently has high requirements for real-time and reliability. For industry, Intel is relying on leading technologies and products, driven by customer scenarios, to create edge computing + artificial intelligence solutions to help customers solve practical problems. In the future, Intel is willing to continue to cooperate with partners to develop technologies and solutions that meet the needs of users in the domestic market and help the development of China's industrial interconnection industry

Copyright © 2011 JIN SHI