“With the emergence of low latency requirements for interactive, cloud gaming and so on, we need to rethink the definition of the system. The first thing that comes to mind is whether we should think of them as a system. The advantage of thinking of them as a system is that it helps us to go back to the origin of the appeal and rethink how to solve the problem. “
The following content is compiled from an interview with LiveVideoStack and Huang Ting.
LiveVideoStack: Please introduce yourself and your recent technology trends and products.
** Huang Ting: ** on the one hand, personal interest, on the other hand, work needs. I tend to follow experts in a particular field on Linkedin to get the information I want, both technical and product. The sources of experts mainly include: the authors of some papers and the contributors of some technical sharing. They can also meet lots of interesting people and the latest product and technology trends. I focus on a variety of technologies, from API design to architecture design and related technologies in the video field. Recently, DUE to work reasons, I have paid more attention to some technologies of RTC.
LiveVideoStack: What are the most memorable projects you have worked on since joining Huawei? Can you share a story or two with LiveVideoStack readers?
** Huang Ting: the one-month recruitment in Ireland in 2017 was very unforgettable for me. During this period, I focused more on people than on things. This work experience has enabled me to get in touch with many excellent people. One common trait of them is that they are willing to share technology and can explain a thing in depth and simply.
At the same time, there is a very mature offline communication platform Meetup and a very good technical communication atmosphere. There are event organizers who are willing to engage in public service, event sponsors (usually companies) provide venues, and participants include university teachers, company technical experts, job seekers and knowledgers. Everyone gets something out of it.
LiveVideoStack: It has been 15 years since you joined Huawei and you have never left. What attracts you most about Huawei, and how has the company changed in your eyes during this period?
** Huawei is like an old wine, the more it tastes, the more it tastes. Huawei has a complete set of management modes and systems for technologies, projects and talents, which is worth our careful appreciation. The other is huawei research and so on. Although I have worked in Huawei for 15 years, EVERY time I hear some very interesting technical exploration directions.
LiveVideoStack: What are some of the most impressive challenges you’ve encountered in exploring video transmission and related solutions?
** HUANG Ting: ** Although I have been engaged in the video field and the products I have made seem to be the same, there are actually many essential differences, which I think is the biggest difficulty. Because once the use of conventional thinking to do a new direction, it is easy to go astray, the more effort may be more wrong. This is what I hope to inspire you with this share.
LiveVideoStack: What has inspired you the most in your technological advancement?
** Huang Ting: ** At each stage, there is a person who is very helpful to us. I think my mentor is the one who has the most influence on me technically. When I just entered the office for a year, there was a conversation, he told me that technical people to do a sweeping monk. It is precisely because of these words that I can calm down to do technical research in my impetuous age. In the face of many internal team and organizational adjustments, I can stay true to my original aspiration, always keep curious in the technical field and insist on doing valuable things.
LiveVideoStack: From the perspective of a pure audio and video technician, how do you understand the value proposition of the new media network proposed by Huawei Cloud and its impact on the industry ecology?
** Huang Ting: ** Broadcasting, transmission and computing (media processing, transcoding, etc.) are relatively independent systems. The advantage of independence is that it can evolve and develop independently without interference and is easy to use. The disadvantage is that there is some loss in efficiency.
With the emergence of low latency requirements such as interactive and cloud games, we need to rethink the definition of the system. The first thing we think about is whether we should treat them as a system. The advantage of treating them as a system is that it helps us to go back to the origin of the appeal and rethink how to solve the problem.
Based on this thinking, we hope to put the computation-intensive graphics processing algorithm originally run on the terminal into the edge or cloud through the innovation of edge-cloud collaborative architecture. This essentially breaks the boundary of the original three systems and provides a new way of video application development. For app developers, it eliminates the need to adapt to multiple terminals and allows product managers to reinvent the video experience to be more competitive. We hope that through our technological innovation, we can enrich everyone’s technological choice.
LiveVideoStack: What are the characteristics of a scalable, reliable, cost-effective video delivery solution?
Huang Ting: ** This is an architectural issue. The reason why the Internet-based Web ecosystem thrives is that it has a scalable, reliable, and cost-effective REST-style architecture. Let’s understand what the REST architecture style has in common:
1. Client-server architecture
2. No state
3. Can be cached
4. Unified interfaces
5. Hierarchical systems
6. Code-on-demand
Recommended here:
Architectural Styles and the Design of Network-Based Software Architectures by ROY T. FIELDING
If the transmission of video can be designed based on such an architecture, the same quality attributes can be obtained. However, more and more services with high real-time requirements appear, which brings challenges to the video transmission mode based on THE REST style architecture (the latency of HTTP, CDN, and E2E is difficult to meet the 100-millisecond delay requirements).
At present, a large number of services with high real-time requirements adopt THE RTC technology based on UDP to transmit video. Since there is no consensus on a unified architecture style similar to REST to meet the needs of scalable, reliable, and cost-effective video transmission, RTC service providers generally provide services in a closed way, that is, provide client SDK. Each uses its own architecture to achieve scalable, reliable, cost-effective video transmission.
WebRTC provides an open source reference implementation that can meet certain open transport requirements. But because its transport standards leave a lot of room for improvement, and it relies on IETF for standardization, it is not evolving as fast as the closed approach. Here is to see the actual appeal of each application to choose.
LiveVideoStack: In terms of providing better video transmission solutions, what other ideas do you think are new and interesting besides Huawei cloud?
** Huang Ting: ** Recently, there are some consumer products that use millimeter wave technology to realize wireless projection screen and HDMI wireless docking. I think they are very novel, and I plan to buy them and play with them.
LiveVideoStack: What do you think of the statement that “cost and technical complexity often face a trade-off in video service quality” **?
** Huang Ting: ** This is the same as the design of the architecture. SLA wants to make it higher, so it may have to pay more cost, for example, there will be more redundancy. I think it is necessary to return to the product positioning itself in the end. It would certainly be better to be flexible in architecture and algorithm design, but this also depends on the technical stack capabilities of each team and what the important issues are at hand. Recently, we are also seeking some human factors engineering experts to discuss the topic of video service quality, hoping to make some new discoveries and breakthroughs in the definition of video service quality.
LiveVideoStack: What is your new understanding of audio and video services & technologies in the post-COVID-19 era?
Huang Ting: In the post-COVID-19 era, some changes may take place gradually, and I think they will be family-centered.
1. More and more fixed scenarios are used, because people work or play at home (with high security);
2. More and more large-screen applications allow better displays;
3. The introduction of new display and acquisition devices will be accelerated, because they can be preferentially used in fixed scenes and require fewer problems to be solved compared with mobile scenes;
4. In the home scenario, heavy entertainment and interactive communication needs will gradually emerge. So the post-pandemic era will accelerate the development of multimodal media technologies, not just by digitizing audio and video to meet human needs. It also includes digitizing human touch, smell and taste to meet people’s heavy entertainment and office needs.
LiveVideoStack: What are the current problems you are solving and what are your next development goals?
** Huang Ting: ** is the three value propositions I mentioned in my last share:
1. Low latency, full interconnection, large-scale real-time audio and video distribution;
2. High-throughput, immersive new media transmission;
3. Collaborative innovation of end, edge and cloud, flexible definition of media processing pipeline.
LiveVideoStack: You have mentioned three major challenges in the future of audio and video transmission in LiveVideoStackCon Beijing station: multiple services, high requirements and fast development. What time dimension does the “future” you mentioned here refer to? In order to meet these challenges, what specific preparations has Huawei cloud made and what product projects have been launched?
** Ting Huang: The three challenges mentioned in ** share are already visible, and I think it’s time to start making technical preparations. We have specific projects to track and implement these three challenges. Huawei has a whole set of process support from research to commercialization.
** Multiple services: ** In order to cope with the increasing number of different types of services on the edge, our newly constructed products such as RTC have been built based on Huawei’s edge cloud products, IEF and IEC to solve the problem of smokestack resource management and greatly improve the operation and maintenance efficiency of products.
** Demanding: ** It can be seen that Huawei Cloud has explored new immersive media directions such as VR and free perspective, and has been able to provide stable products and solutions.
** Fast development: ** Now we have completed the first step of feasibility verification with Douyu. In the future, we will carry out design and research on improving openness and solve the problems of resource isolation and mutual non-influence of algorithms of different tenants.
LiveVideoStack: How do you see the development of video delivery today and the trend of the next ten years?
**5G and new wearable devices are several variables that affect the development of video transmission. I am looking forward to what video technologies will be applied in the Tokyo Olympic Games and Beijing Winter Olympics in the next two years to surprise us. AR and VR are the next step in the evolution of video and even the entire computing and terminal industry.