On March 21, Ali Cloud Summit was officially held in Beijing. The edge Computing special session invited industry experts such as China Mobile, Ali Cloud, Intel and many pioneers in the edge computing industry to share with the audience the layout and the latest technology exploration and practice of edge computing in MEC, IoT, video, edge AI and other directions. It has attracted strong attention from operators, hardware and chip manufacturers, edge computing applications and many other on-site audiences.
5 g era at all things will soon become a reality, put forward new requirements, but for calculating structure need to be more low latency, bandwidth, high concurrency and localization, the cloud, the collaborative computing will may not be able to meet the computing needs of new scene, so the cloud, edge, with three TiXie have become the best solution with high efficiency, real-time and security features on the edge of computing, It will be one of the indispensable infrastructures in the future.
Dr. Wang Jun, China Mobile: Let edge computing be everywhere!
First of all, Dr. Wang Jun from The Technology department of China Mobile shared the exploration of China Mobile in the field of edge computing and his vision of the future of 5G+ edge computing. He believed that in the past decade, the network and cloud have been evolving independently without being visible to each other. And edge computing, inserts “edge” in the middle of “tube”, is the first handshake between cloud and network, at the same time, this is ali Cloud summit ten years to start a key station — cloud capacity to the edge of the expansion.
5 g era and edge of the calculation, the relationship between the characteristics of 5 g large bandwidth, low latency brought up many of the emerging industries, emerging industries and provide edge calculation with much development space, so from that perspective, the edge of computing destined to xing, due to 5 g power 5 g change society together, at the same time to provide better services for you.
China mobile will undertake to build “connect + computing” new converged infrastructure, the role of information transmission and processors, based on the network/IaaS infrastructure, to provide full stack open service mode, will be deployed at the edge of the computing routinization and standardization step by step, at the same time through from the chip, the server to micro data centers all hardware industry chain, realize the authors efficiency.
At present, China Mobile and Ali Cloud have carried out cooperation in edge computing, giving full play to their respective strengths and providing a good carrying capacity of “traffic + ecology” based on “basic resources + operation and maintenance”. It is reported that the two sides have jointly carried out the pilot Edge AI and Edge vCDN in Zhejiang, and will expand the scale of Edge computing cooperation in the future, jointly explore and test scenarios such as industrial Internet and new retail, and jointly incubate Edge applications. “Of all things, hope is the best! I believe that through the joint efforts of upstream and downstream colleagues, edge computing will certainly blend into our lives in the next two or three years, just like 4G changes our lives now.” Wang jun said.
Ali Yunyang Jingyu: Evolution of Ali Yunens from 1.0 to 2.0
After the arrival of 5G and edge computing, cloud and end-to-end two-system cooperation will develop towards cloud and end-to-end three-system cooperation. Undoubtedly, this is a new form of future architecture. Yang Jingyu called the whole process a huge revolution of computing architecture. Ali Cloud edge computing will layer down to the user’s nearest kilometer.
In the era of Edge Node service ENS 1.0, Ali Cloud activated the potential of edge infrastructure covering the world, advancing computing to the 100km edge for the first time and establishing a standard, secure, multi-tenant edge infrastructure base. At the same time, this is the first time that Ali Cloud lightened the Flying technology to a large number of edge nodes, with the help of flying ten years of technology precipitation, to achieve excellent performance, stable and reliable edge computing capacity. But Yang Jingyu thinks this is not just a starting point, edge computing should not just be a simple migration of traditional IDC.
Today, we welcome ENS 2.0. “In the 2.0 era, we will break the fragmentation and make ENS data and capabilities fully connected to the user’s operation and maintenance system,” Yang said. Through the FORM of API, more capabilities are released, so that edge computing applications become simpler, the platform to do the bottom things, the user’s operation and maintenance system is lighter. One-button deployment, one-button upgrade, one-button capacity expansion, automatic alarm response, upstream and downstream interaction to ensure the robustness of the operating system. To be more flexible and flexible in payment models, to be truly asset-light operations, rapid deployment, pay on demand, pay now, don’t need to lock up resources for a year.”
Alibaba Cloud ENS evolution from 1.0 to 2.0
Three-body cloud Li Xingyuan: Practice of audio and video cloud evolution under edge Computing
Trisolaran cloud is a global real-time audio and video technology innovation company. Li Xingyuan, technical director of Trisolaran Cloud, shared its industrial applications in the fields of entertainment and social networking, education and training, video customer service and technology practice of edge computing. Li Xingyuan said that audio and video communication field for real-time, transmission quality to go to very high, which is very big pressure on the core machine room, early often receive alarm load, at the same time, the cost of this program is very high. With the growth of services and the increasing requirements on interactivity and sound quality in business scenarios, it is difficult for traditional protocols and scheduling schemes to meet the demands of large-scale users for audio and video services.
In order to solve the above problems, Trisolaran cloud adopts a series of combination measures such as transmission protocol optimization, computing power improvement (hardware accelerated GPU) and center sinking. According to reports, at present, Trisaran cloud has deployed more than 100 edge nodes, bearing 50% of user traffic, bandwidth cost decreased by 30%, core load decreased by about 40%, basically received little alarm of machine room load.
Li Xing original said: try doing now will be the three body cloud computing resources down on the edge of the core node, on the edge, transcoding, recording, video rivers slice, work done, under this scenario, a lot of video stream is no longer need to do exchange center node, can effectively reduce the cost of consumption, at the same time, also can enjoy better go quickly on the edge of computing power.
Expanding class cloud Chen Yongji: Edge computing solves the problem of flexible deployment, nearby push and pull flow and cost saving
Audio and video technology based on the online education industry service extension class cloud also the site to share the margin calculation in its application on business scenarios, Chen Yongji chief technology officer at the scene said: large-scale online education industry, even the wheat, such as low latency requirements at any time, edge cloud calculation for its class solves the flexible deployment, push-pull flow nearby, and the problems of cost savings.
Talking about the cooperation with Ali Cloud, Chen Yongji said: “After joining Ali Cloud ENS, first of all, we can deploy very flexibly, dynamically increase the number of servers according to the current demand, and the startup speed is in the level of minutes. It can also predict in advance according to big data, so as to be completely unattended. Second, if we want to transmit the flow of quality do well enough, in addition to using WebRTC and the research of algorithm, the loss is another way to put the server deployment in the home, if in accordance with the traditional methods, we communicated with different operators, open room, such as testing, access to the process is complicated, ENS good solve the problem, The network delay between users and edge nodes is as low as 5ms, realizing a global efficient transmission network. In the end, it gives us a lot of operational savings, as well as a lot of capacity on edge nodes and a lot of cost savings on machines and bandwidth.”
Innovative applications based on Intel platform technologies
Intel Li Jinwen talked about the innovative application based on Intel platform technology: 5G is characterized by large bandwidth, high reliability and interconnection of all things. At the same time, the number of interconnection will be hundreds of billions, and the amount of data brought by such a large connection will also be huge. If all the data are put in the center, the pressure from the center, bandwidth and cost will be very great, which is why we need edge computing.
Intel and AliYun have conducted closer cooperation in data center – cloud computing, big data, AI, intelligent manufacturing, system software and other aspects. At the same time, the video cloud market has a large number of scenes that can be applied jointly with edge computing, such as sports live broadcast, media analysis, realistic immersive experience, etc. “Everything is smart and connected, and we believe edge computing will play a role in more scenarios than 5G smart cities, video cloud markets and intelligent manufacturing,” Li said.
Aliyun Liu Haoran: Link IoT Edge – Make IoT intelligence one step closer
In recent years, the application of the Internet of Things has gradually entered into production and life, and the combination of the Internet of Things and cloud computing has become closer. However, the Internet of Things industry is facing many pain points, such as: How can the existing equipment be transformed into the cloud quickly? How to deal with service interruption when the network is disconnected? On the original massive data, the cloud cost is high, low latency fast response, data privacy and other security issues, how to solve these pain points, Ali Cloud product expert Liu Haoran in “Link IoT Edge – Make IoT Intelligence one step Closer” for us to answer.
Link IoT Edge sinks the cloud capability to the Edge and uses the cloud capability to solve local problems. It focuses on providing device connection services, device management services and device linkage services under the security protection of the Internet of Things, enabling all kinds of Internet of things scenarios and industry developers.
In this architecture, Link IoT Edge provides integrated operation and maintenance, end-to-end development tools, scenario applications, and device drivers. First of all, Link IoT Edge is a cloud and Edge integrated product, including cloud memory service and Edge SIDE SDK. It integrates more than 10 cloud products such as security, storage and computing, making it very convenient to access cloud services on the Edge. Thirdly, more than 10 scenarios have been implemented, including Internet of things fields such as parks, human settlements and industry. Finally, the IoT scene is very fragmented and ecological is very important. Currently, Link IoT Edge supports more than 30 kinds of hardware, including mainstream platforms covering x86, ARMv7, ARM V8 and other mainstream chip architectures.
In the future, Link IoT Edge will strengthen the ecosystem of hardware and software, accelerate the integration of Edge services, continuously improve the efficiency of business operation, and help more enterprises achieve efficient innovation.
5G era of everything Internet has arrived, ali Cloud edge computing and industry ecology, application collision what spark? We’ll see.
The original link
This article is the original content of the cloud habitat community, shall not be reproduced without permission.