Audio and video technology in the live, chat, games, and other areas of the extensive entertainment applications, there are already a lot of depth in the field of Internet education has also become the standard, the financial industry of video recording, online business hall, the insurance industry in the video to open an account, to open early will, with the deepening of industry informatization, the degree of digital audio and video technology is more and more perfect. LiveVideoStack specially interviewed Liao Nianbo, vice president of Technology of Struct technology, and asked him to tell his understanding of the background architecture, operation and audio and video industry of video cloud PaaS service from the perspective of a veteran of technology in the background.

The expert information

Liao Nianbo is the vice President of Technology and the head of background technology of Jiku Technology. He has 16 years of experience in Internet massive service architecture and technical operation. Previously, I worked as the technical director of Tencent QQ and Tencent Video, responsible for the basic background team and AI technical team.

The following is compiled from an interview with LiveVideoStack and Liao Nianbo.

01 Is currently responsible for the main business in STRUCt

I am in charge of the background team of real-time audio and video cloud PaaS service in Jike Technology. Based on our PaaS service, customers can quickly achieve:

1. Real-time audio and video communication of multiple people, such as video conference, online education, live broadcast with microphone, game opening black, etc.

2. Large live broadcasts, such as shows and concerts.

Customers do not need to care about the encoding and decoding, streaming media transmission, server development and operation and maintenance.

What are the reasons for joining STRUCt? Are there any discrepancies between previous work experience and the work in THE STRUCTURED technology?

There are three main reasons why I joined STRUCt.

1. The architecture and operation experience I have accumulated in Tencent can be put to good use and give full play to my strengths.

2. Compared with the toC/ localization products I have experienced before, the toB PaaS service provided by The toB technology is a higher challenge for me to practice the shooting range better. For example, multi-tenancy requires a more resilient architecture where resources are shared and isolated. Users from all over the world demand that our services be decentralized and globally distributed; Real-time communication requires the pursuit of low delay.

3. Technology is a company based on technology. Technology is our primary competitiveness and can significantly bring value to customers and the society, which is in line with my original intention as a technician.

Speaking of similarities and differences, I think these two careers are similar in terms of technology and methodology. The difference, in addition to the specific technical challenges mentioned in point 2 above, is that it is clear that even though technology is a growing company, I feel there is more to do and more room to play.

In fact, this is not just me, from ordinary employees to senior architects or management cadres, we all have this experience. It doesn’t happen everywhere that a brilliant young man, just a year or two out of school, has the opportunity to build an important system. In this process, you will feel a great sense of achievement, you will feel the obvious growth and transformation of yourself, of course, there will be a lot of difficulties, pain, anxiety, but also happiness.

During my 20 years of working in the Internet industry, I was particularly impressed by the technical difficulties

Memories are always simpler and easier than experiences. Looking back now, it seems that no specific technical point or major project event is particularly memorable. Although along the way, I often struggled with some technical difficulties and brain, such as how to abstraction and decoupling to make the architecture more clear and extensible, how to achieve fast and reliable synchronization of data around the world, and how to achieve higher performance network transmission…… But these jobs have become routine, albeit uneventful.

However, in recent years, I feel more and more deeply that the technical operation ability can be highly competitive, which is a difficult problem that can reflect the team level. Teams with weak technical operation ability either put out the fire all day long, or sort out and rectify the symptoms rather than the root cause. With the growth of business and the evolution of demand, the complexity of work increases exponentially, and the system stability cannot be sustained.

In contrast, a team with strong technical operation capability has several characteristics.

1. The work is done in a comprehensive way, without any obvious health care factors. For example, the impact of Agile software development, where business features are quickly introduced without logging/monitoring/operations tools and no effort is put in to fix them, can snowball. Agile software development is great, but some health factors must be added in time.

2. Operational capabilities are designed into systems from the start with as much (or even higher) priority as business features. Good systems are “designed”, and while the methodology of continuous iteration and making products grow like coral is great, this applies primarily to business features/product experiences, and the underlying framework needs to be designed from the start, otherwise every iteration will break your bones.

3. Use software engineering/tools instead of processes and relying on people awareness for team synergy and capacity accumulation. DevOps tools take a requirement from proposal to release like a factory line; Automated diagnostic tools allow operations personnel, with limited information, to retrieve the first scene from extremely complex systems and get to the heart of the problem. Robust design allows the system to withstand falls like a cow rather than a pet. Chaos engineering/manometry pushes online systems into a corner and lets actual performance speak for itself.

04 Technical challenges encountered by audio and video communication technology

Thanks to the rapid development of the Internet, basic science and infrastructure around the world, audio and video communication technologies are getting better and better. More and more people are using them, covering more and more fields, and users have higher expectations. But the basic technical challenge remains the same: how to ensure stable, low-latency data transmission in a world of complex networks.

From the first kilometer on the user side to the public network relay between hundreds of countries and regions around the world, the physical distance is far away and there are many links, which can be regarded as a complex chaotic system. And getting steady, real-time audio and video data is like getting a letter to Garcia through a barrage of gunfire.

Our team developed Massive Serial Data Network (MSDN) through Massive service practice accumulation, combined with SDN technology and backed by basic multi-cloud business, which can automatically tolerate faults, intelligently select the optimal transmission path, and respond/automatically recover in case of line failure. Let users get higher network quality.

According to the business needs of different customers, the company provides “expressway” and “professional track” of data transmission respectively, which is to say, low-latency Live (L3) products and WebRTC ultra-low delay technology.

L3 reduces the transmission delay from 3-5 seconds of traditional CDN live broadcast to 1 second, effectively improving problems such as high delay, poor resistance of weak network and inconsistent content in scenes of “online education, e-commerce live broadcast, sports live broadcast and show live broadcast”. At the same time, it also breaks the dilemma of only choosing between CDN and RTC in the market, and helps customers take into account cost and performance, intelligent scheduling and balanced resource allocation.

With the support of self-developed audio and video engine and MSDN network, we have recently made a new breakthrough in low-delay data transmission, realizing the end-to-end ultra-low sensory delay of 70ms from “information collection -> processing -> coding -> transmission -> decoding -> rendering”. Under the time delay of 70ms, the sensory delay of human body is almost zero, which can be applied in scenes with higher requirements for real-time interaction feedback.

This technology has been applied to the scene of online karaoke, and the “Online KTV real-time chorus solution” has been launched. The system has solved the problem that chorus can’t align with chorus in real time in the process of online karaoke. We are the first technology service provider in the industry to implement the real chorus scene.

Opportunities for audio and video technology to translate into commercial value

Namely structure since entrepreneurship has been holding the idea of “for audio-visual technology to dissolve into the invisible,” our iteration steps, real-time audio and video product at the same time in different level of the cloud communication, introduced a real-time news, low latency, live visual products such as products, AI, visible to the naked eye becomes rich three-dimensional technology system, The idea of “invisibility” is deeply imprinted on each of us. Because not only technology and products are moving forward, production processes and production scenes in more industries are also innovating. Such changes do not mean that audio and video technology should become the leading role in the transformation of the industry, but various industries include audio and video technology in a larger range. So we also chose to use a more structured and hierarchical product matrix to try to integrate more and see some chemical reactions.

Av technical support, in addition to the bottom in order to meet the needs of more ability to quickly get audio and video, we will also PaaS products, penetrate the specific business scenarios, providing low code, extensibility and elastic telescopic aPaaS mode solutions, let the customer can be faster and more cheaply audio and video products.

In February this year, we launched RoomKit, the industry’s first low-code interactive platform for the whole industry. Through the complete encapsulation of business scene capabilities, interactive rooms can be built in zero code. Even customers without a technical team can access and bring products online through Roomkit’s functional visualization. In addition, we also provide Talkline, Xiaoyi Gang and other App products.

Why do you do it this way? We were born in the Internet industry, where the degree of digitalization is relatively perfect. Audio and video technologies have long been extensively applied in the pan-entertainment fields such as live broadcasting, chat and games, and all kinds of gameplay can be said to be rich in imagination. In addition, audio and video have become standard in the field of Internet education. But to say that in traditional industries, because of the difference in the progress of enterprise informatization, we are feeling the change before and after, this phenomenon is actually quite interesting. For example, the traditional education and training industry uses Roomkit for rapid online transformation. In the financial industry, video recording and online business hall, video account opening in the insurance industry is used for morning meetings, and small art band is used for school online exams and remote enrollment. With the continuous deepening of industry information, I believe that such examples will only be more and more.

I think in the whole process of integration with the industry, the key to turning technology into business value is “patience”. We expect to go deep into the industry to produce chemical reactions, rather than quickly push the physical reaction, so we must try every means to stand in the customer’s perspective to try to create value for their industry, from the industry’s long-term benefit from information development, this is the healthiest. But having chosen to become a “partner” in the early stages of the industry’s transformation, it is naturally more bearable for a moment of loneliness.

Get the most out of tech conferences

Because time is limited, come and go quickly 🙂

But seeing more and more companies/people join or pay attention to real-time audio and video technology, talent advantages and fierce competition can make the whole industry do better.

In addition to learning from others, attending the technical conference is also a process of systematic thinking and summary for myself, which makes me gain a lot.

Editor: Cindy Chen

Activity recommended

Of the world’s leading real-time audio and video cloud service providers that compose technology on May 29 (Saturday) to be held in Beijing union volcanic engine “extensive social audio and video entertainment technology practice salon”, specially invited to think at infinite (thorn live), director of research and development, the structure of science and technology solutions senior architects, volcanic engine solutions, senior advisor to the three guests, Respectively from the experience to upgrade and upgrading of technology to promote the content of the live entertainment, Shared experience scenario innovation, RTC service experience optimization, such as audio and video business link growth all dimensions to share best practices, checking of actual combat experience, chat about technology trends and future play, / pan entertainment social scene of audio and video technology innovation interested friends, can scan qr code below poster