This article is the development trend of digital services in "China Intelligence Observation" "Enterprise Digital Service Supply and Demand Market". With the rise of China's power, large amounts of funds have poured into edge computing, and the application scenarios are complex and changeable. Will edge computing become the fourth wave after cloud computing? The author explores with you.
——Haibi Research Institute
February 21, 2022
is like a magical octopus, edge computing allows application agility, real-time and other problems to be solved easily.
header image source: Photo Network
For self-driving cars, lightning-fast and 100% accurate decisions must be made. If a child runs across the road, the car cannot risk even the slightest delay before brakes.
However, cloud computing is connected to a centralized data center, process the data in a remote data center, and then returns the result to the car. Every millisecond is precious, and this method is obviously unacceptable.
edge computing is to process data on the "edge" of the network. It eliminates the need to connect to remote data centers and is therefore faster than cloud computing. Research shows that edge computing can increase processing speed by 30 times! The most critical thing about
is that edge computing makes "impossible". Without edge computing, the commercialization of autonomous vehicles, the Internet of Things, AR and 5G will never be able to start.
However, if you are pursuing 10x or 20x profits like Amazon cloud computing or Alibaba Cloud, then now you should focus on the next opportunity for computing to change, which is edge computing.
Amazon cloud computing, Salesforce, Microsoft, or Alibaba Cloud, Tencent Cloud, Huawei Cloud, including China Mobile, China Telecom, and China Unicom, have all stepped on the pace of computing development and are opening up the market in the field of edge computing.
1. Capital trends: A large amount of funds poured into edge computing
Edge computing is the fourth climax of computing?
Although edge computing is not as popular and popular as technologies such as artificial intelligence (AI), 5G, augmented reality (AR) and the Internet of Things (IoT), large amounts of funds have poured into edge computing and the application prospects are becoming clearer.
Because, the industry firmly believes that after mainframes, personal computers, the Internet and cloud computing, edge computing is considered to be another revolution to change the IT computing industry and will create new trillion-level corporate giants.
GrandView Research expects the edge computing market to grow at a rate of 54% per year and will reach nearly US$29 billion by 2025.
Data from global research firm AnalysysMason shows that 30% of enterprise IT budgets will be used for edge computing in the next three years.
Different institutions have analyzed the development of edge computing from different angles. The
Gartner research report predicts that by 2025, the proportion of data generated by enterprises processed outside of "traditional" centralized data centers or cloud storage will jump from 10% in 2018 to 75%.
Forrester also believes that the edge cloud service market will grow at more than 50% in the future. 2020 will be the "breakthrough year for edge computing", and edge computing may surpass cloud computing after that.
In short, edge computing can easily become one of the latest and best investment opportunities in the digital age we live in. But it won't be a "winner-takes-all" market. As edge computing flourishes, many companies that grow 10 times or better will be born.
The true or false of edge computing?
In the development of edge computing, there has been controversy over real and false edge computing. This is actually not very meaningful for the development of edge computing.
edge computing refers to processing data on devices such as smartphones. Unlike cloud computing, cloud computing processes data in remote, remote data centers, edge computing enables devices to perform part or all of the data processing immediately when data is collected.
Because smart devices are becoming more and more powerful, devices no longer need to send every small piece of data (whether useful or not) to the cloud, and can handle more data and tasks.
For example, the office security camera collects all data overnight, with the vast majority of video data showing empty corridors and rooms. Sending all this data to the cloud is obviously a waste of bandwidth. However, an AI-equipped security camera can immediately analyze images, detect abnormal activities and alert them, and can quickly implement security monitoring functions.
Many experts, for example, octopus uses "edge computing" to solve practical problems. As one of the most IQ animals among invertebrates, octopus has a huge number of neurons, but 60% are distributed on the eight legs (wrist foot) of the octopus, but only 40% of the brain. In other words, octopus uses its legs to think and solve problems nearby.
Cloud computing and edge computing are complementary.
In the future, will edge computing be better than cloud computing? Actually, it is not the case!
Cloud computing is the interaction between people and computing devices, while edge computing is the interaction between devices and finally serves people indirectly. Edge computing can process a large amount of instant data, while cloud computing can finally access the history or processing results of these instant data and perform summary analysis.
If cloud computing is the brain of an octopus, then edge computing is the antennae of an octopus. Most of the responses of the antennae to external stimuli are instinctive, and the results of these continuous stimuli will eventually be collected into the brain, and then serve as a basis for decision-making in the subsequent behavior of the antennae.
From this point of view, cloud computing and edge computing are a symbiotic and complementary relationship. There will be no problem of who will replace who will be there now and in the future, but who will have more advantages in which computing and who will be more suitable for which scenarios.
2. Application scenarios: Edge computing is facing the development of complex and changeable
cloud computing, IOT, AI, 5G and other technologies, giving wings to the development of edge computing.
In edge computing, you can find the shadow of these technologies everywhere.
Cloud computing provides computing, storage and network services to enterprises. Object storage services such as AmazonS3, AzureStorage, and GoogleCloud Storage become storage devices for content used by managed workloads.
AmazonCloudfront, AzureCDN, GoogleCloud CDN, etc. CDNs have become logical extensions of object storage, used to distribute and cache content across edge site networks.
The rise of the Industrial Internet of Things (IIoT) has led to the introduction of IoT gateways—a dedicated device that converts protocols used by local devices into cloud protocols. The IoT gateway also acts as a data aggregator, combining and multiplexing telemetry streams from multiple devices and filtering them before streaming to the cloud.
Recently, artificial intelligence has become a key component of IIoT. By deploying deep learning models at the edge, organizations are able to perform inference in real time. Predictive maintenance – a method to detect equipment and mechanical failures before actual interruption requires faster turnaround times. IIoT customers want to run AI models locally by keeping them closer to the devices that act as data sources.
infrastructure running on telecom operator facilities connected over 5G networks has low latency. Telecom operators such as China Telecom, China Mobile and China Unicom are turning to the multi-tenant hosting infrastructure layer to bridge the gap between the cloud and end users. Amazon, Google, IBM and Microsoft, as well as cloud providers such as Alibaba Cloud, Tencent Cloud, and Huawei Cloud are working with telecom companies to introduce some hosting services to 5G-based edge sites.
edge computing covers the entire range from device to cloud, with a large enough range to meet the development needs of different enterprises.
Micro-edge is the latest incarnation of the edge computing layer. When a microcontroller is able to run a TinyMLAI model, it qualifies as a micro-edge computing device. In this use case, sensors connected to the microcontroller generate deep learning models for inferenced telemetry streams. Unlike other solutions where microcontrollers collect telemetry data and introduce edge computing layers, this type of edge runs in the context of microcontrollers and microprocessors.
Mini Edge , for example, based on single-board computers based on ARM64 and AMD64 architectures. It is usually powered by AI accelerators to speed up reasoning.It is also able to run mature operating systems such as Linux or Microsoft Windows. MiniEdge comes with a software stack associated with an AI accelerator. These types of edge devices are ideal for protocol conversion, data aggregation, and AI inference. The edge deployment model in
represents a cheap computer cluster running at the edge computing layer. Computing clusters are powered by internal graphics processing units (GPUs), field programmable gate arrays (FPGAs), vision processing units (VPUs), or application-specific integrated circuits (ASICs). Cluster managers like Kubernetes are used to orchestrate workloads and resources in a cluster.
Heavy edge computing device is usually a hyperconverged infrastructure (HCI) device that runs within an enterprise data center. It comes with an all-in-one hardware and software stack that is usually managed by a vendor. Severe edges require only power and network resources available in environments such as enterprise data centers.
AWSSnowball Edge, AzureStack Edge, NVIDIAEGX A100, and NutanixAcropolis are some examples of heavy edges.
Multi-access edge computing (MEC) transfers traffic and service computing from centralized cloud to network edge, closer to customers. With 5G becoming a reality, MEC is becoming an intermediary between public cloud consumers and providers.
AWSWavelength, AzureEdge Zones, and GoogleGlobal Mobile Edge Cloud powered by Anthos are examples of MEC.
Edge Cloud performs what CDN does for static content, but for dynamic workloads. It allows the distribution of components of the application between multiple endpoints to reduce the latency involved in the round trip process.
edge cloud relies on modern application development paradigms such as containers and microservices to distribute workloads. The static content and stateless components of the application are replicated and cached throughout the global network. Edge cloud providers may support AI acceleration as an optional feature. Since it is provided as a hosting service, customers do not have to deal with hardware and software maintenance.
edge computing and ecosystem definitions are rapidly evolving to meet the needs of enterprise customers.
3. China's power: the new rise of the edge computing market
China Software Network found that among Chinese edge computing companies, it is mainly divided into five categories:
one is cloud computing manufacturer. is based on its own powerful IaaS and PaaS. Cloud computing enterprises use distributed technology to expand and extend computing power and storage outward, covering edge computing from top to bottom and from the inside out, while the control center is still firmly in the central node.
Typical enterprises include Microsoft, AWS, Google, as well as giants such as Alibaba Cloud, Tencent Cloud, and Huawei Cloud.
For example, Huawei's all-round edge computing players, the heavy edge is mainly concentrated in MEC products that cover 5G core network, 5G base station, 5GUPF edge network elements, MEC edge MEP/MEPM and other MEC edge MEP/MEPM.
light edge field has launched IoT gateways such as Huawei AR502H, providing SDK to achieve flexible calls to computing, storage, and network resources.
Alibaba Cloud relies on edge technology, ENS is built based on CDN layout, and is also planned to be built based on operator edge nodes and networks. Light edge technology OpenYurt is an open source IoT device light edge dock based on Kubernetes, which can achieve light edge application sinking with Alibaba Cloud.
is a traditional telecom operator. Compared with other players, the resource advantages and basic layout of have won at the starting line in the competition for edge computing.
Telecom operators already have contacts with corporate customers, integrating corporate communications, basic networks, personal/group businesses, and packaging edge computing businesses will create quite strong attractiveness and market competitiveness.
operators have begun to deploy mobile edge computing (MEC). China Mobile has carried out pilot projects on various MEC applications in more than 20 cities and cities in 10 provinces. As early as January 2018, China Mobile Zhejiang Company and Huawei took the lead in laying out MEC technology to further promote the network to achieve ultra-low latency and better experience, and create a future artificial intelligence network.
China Telecom cooperates with CDN companies to serve as an extension of existing centralized CDNs through the deployment of MEC edge CDNs, and at the same time serve multiple network users.
Deutsche Telekom uses edge computing in improving connectivity, digital transformation and advancing better network performance for 5G.
is a CDN manufacturer. The core value of CDN (i.e., content distribution network) is to intelligently distribute digital content to nodes closer to users. Relying on edge servers deployed in various places, the central platform's load balancing, content distribution, scheduling and other functional modules enable users to obtain the required content nearby and improve the response speed of users' access.
Its innate edge node attributes, low latency and low bandwidth give it a first-mover advantage in the edge computing market.
Typical examples such as Akamai, as the global leader in CDN, cooperated with IBM in edge computing as early as 2003 to provide edge-based Edge-based services on its WebSphere.
Netsu Technology has also regarded edge computing as its core strategy. It began to build an edge computing network in 2016, gradually launched edge computing microservices in 2017, and will gradually open edge IaaS and PaaS services.
is a chip/terminal device enterprise. edge computing is also a rare opportunity for chip/equipment manufacturers.
In terms of general-purpose chips, ARM has almost completely lost opportunities in the low-power market, and the combination of ARM+ Linux occupies almost the entire smart hardware market.
Due to the rise of edge computing technology, especially on the device side of face recognition, voice recognition capabilities, etc., ARM's advanced chips have begun to be market-oriented and can advantageously support the development of AI.
Huawei HiSilicon chip is now the leader in the video processing industry. From chip to camera, it is not difficult for Huawei, which started with hardware. Especially after the camera becomes a sensor, combined with the capabilities of the cloud platform, its development space and its imagination space are available.
is a traditional transformation hardware manufacturer. hardware companies mainly consider adapting to the overall development of edge computing, developing products with higher performance and lower costs, and at the same time strengthening ecological construction and promoting their own R&D direction to become an international standard and industrial consensus, so that their products can adapt to more partners. Once the industrial cake is bigger, they can naturally make a fortune.
Dell announced its entry into the Internet of Things market as early as 2016, and was the initiator of the edge computing project under the Linux Foundation. EdgexFoundry, an open source project under the Linux Foundation, is committed to developing an edge computing platform with plug-and-play functions. Dell has taken the lead in launching an Edgexfoundry-based edge gateway.
The new DellEMC VxRail satellite nodes reduce the company's comprehensive range of hyperconverged infrastructure (HCI) equipment to the size of a single rack unit, allowing it to adapt to a wide variety of locations.
For specific scenarios of edge computing, some companies will develop more professional products, such as intelligent hardware that integrates AI algorithms, cameras that integrate image recognition and video compression, and Internet of Things modules that integrate storage and communication capabilities.
terminal contacts provide sensors, cameras, robots, on-board equipment, etc. This category includes terminal suppliers such as Hikvision, Dahua Co., Ltd., Advantech Technology, and Aerospace Electronics, as well as base station providers such as Huawei and ZTE.
edge computing is called the "last mile of artificial intelligence". Its advantages such as saving bandwidth, reducing latency, enhancing security and privacy are driving a trillion-level market opportunity.
Technology Media·Enterprise Evolution Science·Strategic Hosting Complex
The author of this article is Haibi Research Institute·Zhao Manman
