//
The application of artificial intelligence represented by ChatGPT is accelerating the change of people's life and work style, and the application of various AI has triggered a new round of productivity revolution in audio and video content. With the generation of massive audio and video content, a large amount of content will be created and stored on the side, and will be calculated and distributed on the side. Wangxin Technology continues to focus on edge cloud computing technology, leverages the advantages of "edge cloud + AI" fusion technology, and actively promotes the development of computing and network integration industries.
LiveVideoStack invited Li Hao, CEO of Wangxin Technology, to tell us about the journey of Wangxin Technology to become an advanced, specialized and special new enterprise, what opportunities and challenges generative AI will bring to Wangxin Technology, and the future direction of Wangxin's business evolution.
LVS: What was the initial opportunity that made you decide to join Wangxin and start a business?
Li Hao: I think entrepreneurship itself is doing the right thing at the right time. During the period of rapid development of the domestic cloud computing industry, I participated in the creation and operation of public clouds, and gained a deeper understanding of customer needs, core technologies, product construction, and service delivery in the cloud computing industry. At that time, I thought that if the future computing is truly cloud-based and within reach, it must not be just some super data centers delivered to users from the perspective of resources, but should be composed of multi-level and heterogeneous computing resources . For complex networks , users can achieve unified scheduling and use through simple methods such as serverless. At that time, it was seen that the edge side closer to users was blank, and there should be opportunities in the field of edge computing.
LVS: Wangxin is a specialized and special enterprise focusing on the field of edge cloud computing. Can you introduce how Wangxin advanced to specialize and specialize?
Li Hao: The development of Wangxin is very in line with the requirements of "specialization, specialization and innovation". The edge cloud computing track we are in is the next market segment of cloud computing. Focus and refinement mean that we can be more advanced in some technologies. polished. At the same time, it is in a leading position in the same industry in terms of innovation ability, market share, mastery of key core technologies, quality and efficiency:
First, market position : Wangxin Technology has a full range of edge computing products, and the number of customers accessing SDK has reached 600 million; Wangxin Technology's edge CDN market share ranks first in the industry. According to the latest IDC report data, Netcent's market share in 2022 will rank third in China's edge public cloud service market;
Second, leading technology : Wangxin Technology has a full-stack edge computing solution, the best technical indicators in the industry, and excellent and innovative edge computing products;
Third, resource advantages : Wangxin Technology has the world's largest and most sinking edge computing network, considerable advantages in bandwidth and computing power, and owns self-developed edge intelligent hardware devices to adapt to various edge scenarios;
Fourth, the innovation team : Wangxin Technology is the earliest manufacturer involved in edge computing in China. It insists on technological innovation. It currently has R&D centers in Beijing, Shenzhen, and Xi'an, and the company's R&D personnel account for 70%.
We will continue to persist in cultivating the edge computing track to provide customers with valuable services.
LVS: AI and AIGC are the current hot topics in science and technology. With the rise of a new round of AI technology storm, what opportunities and challenges will generative AI bring to edge cloud technology?
Li Hao: The rapid development of AI is magnifying the value of data, and at the same time brings massive heavy-duty computing requirements, which are difficult to complete on the terminal, and the edge side will be upgraded from a data cache to an important data storage and computing place. Bringing changes in computing architecture, data and computing power will be redistributed in terminals, edge nodes, and central cloud computer rooms. Changes in computing levels will lead to complex architectures. How to shield these complexities and provide users with simple and insensitive interfaces for data storage, transmission, and computing, while considering the needs of users’ scenarios and the costs of different levels, to do a good job of intelligent allocation, this is It is a big challenge that requires efforts in resource orchestration, demand perception, unified connection, and cost scheduling.
LVS: From the perspective of technology, the deployment of large models on the edge and mobile terminals will become the main trend of future development. Among these, what specific attempts and explorations will Wangxin make?
Li Hao: From the perspective of AI computing power consumption, with the successful implementation of some models, the computing power used in the inference link will be tens or hundreds of times that of the training link. We focus on reducing the total computing power cost of customers, and focus on the edge side. The use of inferential links.
Based on the "cloud (training)-edge (reasoning)-end (request)" full-scenario AI training and reasoning collaboration framework, the network center integrates the heavy computing power of the cloud to support model training, deploys to the edge with one click, and can use a large number of distributed Edge computing power nodes can also reduce data backhaul and greatly reduce reasoning costs. They have been applied in the field of Vincent graphs and some language models. In the field of large model training, we are also trying to effectively utilize distributed computing power nodes.
LVS: As a more subdivided edge cloud computing, some people predict that it will achieve a higher growth rate than the overall cloud computing market. Faced with this situation, what is the future direction of Wangxin's business evolution?
Li Hao: Wangxin currently has its own relatively comprehensive and long-term development plan in terms of business and layout, and the "three-step" business development strategic plan has become increasingly clear.
The first step is to create new potential for sustainable development of customers . At present, based on the scenarios that can generate economic benefits and the actual needs of customers, we use customers to expand the network scale, improve the network level, reduce costs, improve efficiency, and promote healthy expansion of enterprises.
The second step is functional iteration . Gradually enrich business scenarios, promote the improvement and promotion of cloud games, AIGC and other businesses; insist on research and development and coverage of computing power needs; continue to improve the computing power of the network.
The third step is to build an ecology . Promote the superposition of edge network and computing power, connect with high-value edge computing industry scenarios such as intelligent driving and remote control, build an open network platform, and attract developers and partners to jointly build an ecology.
Under the clear strategic planning and layout, we are confident that we can fully realize the technological innovation gene of Wangxin and the current performance results in the next 8-10 years. With the acceleration of AGI, I believe this time may be greatly compressed.
LVS: Which edge computing application scenarios are you optimistic about in the future?
Li Hao: I am very optimistic about the application of edge computing in the field of audio and video. We believe that the audio and video industry will become a basic industry in the future. The development of AI will bring a huge impetus to the field of audio and video. Audio and video are the most abundant information transmission methods. AI will greatly reduce the production cost of video content, greatly enhance the ability of video understanding, and will continue to improve the data of audio and video. If the total amount is increased, it will also bring about an exponential increase in traffic, storage, and computing power requirements. As I just said, we need to upgrade the computing architecture to cope with it. The terminal side is thinner, the core realizes user interaction and audio and video content presentation, and the edge side is thicker to increase storage and computing power to support cloud games, Metaverse and other scenarios. Edge computing will also change from helping business cost reduction to helping business efficiency .
LVS: Entrepreneurship is like a never-ending climb, with obstacles and long roads. From 15 years since its establishment, Wangxin can be said to have completed the transformation process from 0 to 1. What advice or suggestions do you have for those who are just starting out?
Li Hao: Advice and suggestions are out of the question for the time being, and we are still exploring on the road. From our experience, it is not easy to start a business first, and the early stage of idea verification and business model exploration is particularly difficult. The most important thing is to choose a business with the support of vision, think clearly about the direction to cut in, and the value to be realized . To focus . Try to find a group of like-minded partners who have the courage to take risks and the patience to persevere in the trough. In addition, in the current market environment, cash flow is extremely important, especially in the exploration stage of business models, and costs must be well controlled. The satisfaction of realizing value in the process of starting a business is difficult to obtain in other places. The small partners who choose the entrepreneurial path work together.
▲Scan the QR code in the picture or click " Read the original text " ▲
Check out more exciting topics of LiveVideoStackCon 2023 Shanghai Station