Build an XR Collaborative Ecosystem

01 Interactive upgrade, abstract and concrete, immersive experience brings changes to the industry

We believe that whether it is the current term "meta universe" or the new wave of technology, we will call it the evolution of interactive technology. As for interactive technology, it has been making continuous progress in the past hundred years, especially in recent decades. Its evolution is also directional and has rules to follow. What we see is that interaction has gradually changed from one-way to two-way, from abstract information to concrete information.

As a popular understanding, radio and newspaper, as the media, are mainly supported by text content, which is very abstract. At the same time, it is difficult to communicate directly with the author behind the newspaper, which is what we call the difficulty of interaction. Similarly, it is also very difficult to interact with some characters in the novel, because it is one-way. But now with interaction technology, we may wear an XR helmet and interact with any object, environment and person in the virtual world, which also realizes two-way interaction.

So what is the significance or function of realizing effective two-way interaction? In fact, we regard such two-way interaction as the integration of the real world and the virtual world, and this integration has also produced many benefits. The real world is rigid and many things are difficult to change, but the virtual world is flexible. We can customize it at will. When we combine the characteristics of the real world and the virtual world, we find another model to solve the problem.

For example, we want to plan a CBD in a certain place, but we can't accurately predict the impact of the CBD on the surrounding traffic after its completion, which will cause traffic congestion in the surrounding areas after its completion. This situation occurred from time to time in the early days. At this time, if we put the CBD model into the virtual city, and introduce all the traffic elements, including cars, two-wheeled vehicles, pedestrians, etc., and put the corresponding parameters into the simulation. We can estimate where a CBD needs to be adjusted and modified in a certain place, and finally form a model. To sum up, it is to bring the problems of the real world into the virtual world to simulate and find the answers, and then influence the real world again.

The application of combination of virtual and real has been very common in our life and work. For example, in the consumer-end scenario, we often see various digital virtual people, the metauniverse concert, the immersive digital museum, various virtual simulation systems in the enterprise-end scenario, digital twin factories, and so on. At present, the main immersive applications, or visual computing applications, are mainly in these two relatively large scenarios. In summary, the advantages they bring are: in the consumer side, they bring beauty and fun to everyone's life, and in the enterprise side, they bring creativity and productivity.

The presentation of such multiverse applications requires cloud computing. Let's look at the challenges of these applications to cloud computing.

02 Alibaba Cloud, break through the challenges of the industrial chain and open up business cooperation

In our view, this new industrial innovation will affect the whole industry, and all metacosmic applications or visual computing applications can be represented by this model. The metauniverse application developed by developers needs to bring virtual experience to users through cloud computing capabilities and human-computer interface devices. Even if the model looks so simple, it also faces great challenges in the actual implementation process of the industry. The reasons for the challenges come from three aspects: the technology stack of the meta-universe industry is extremely complex, and there are many industries to be affected, as well as many potential users. Finally, it brings two relatively big challenges: technical challenges and commercial cooperation challenges.

Let's look at the technical challenges first.

The technical challenge is mainly the computational challenge. In the past 20 years, everyone has experienced the transformation from PC to mobile phone. The most important change comes from the change of interaction mode. Everyone's interaction has changed from mouse button+display in the PC era to the current multi-touch mode, and even the use of gesture and voice interaction on XR helmets and glasses. These interactions require more computing power, but the evolution speed of computing power of terminals is increasingly tight, and the interactive computing load is gradually moving to the data center.

In the past 20 years, the scale of computing power in the data center has increased by a hundred times. In addition to the incremental changes brought by the increment of Internet users, the rest is the increment of computing brought by the interactive mode. The most significant change is the emergence of heterogeneous accelerated computing in the data center. For example, our common mobile phone voice recognition, shopping recommendation, face payment, AI camera and so on need the assistance of cloud AI computing resources. This is especially true for the XR terminal in the next phase. Let's see what computing requirements are in the XR phase.

Now at the XR stage, the demand for computing is even stronger, even hundreds of times higher than the previous demand.

For example, in the XR scenario, there will be three types of computing: graphics and image computing challenges, AI computing challenges, and streaming media computing challenges. Graphics computing needs to always achieve real-time image generation, AI needs more, real-time dialogue, real-time facial expression changes, real-time emotion recognition and generation, which will face huge incremental computing challenges. In addition, we need another thing in the process of transferring content from the cloud to the user. Communication requires very large bandwidth and very low latency.

In short, the improvement of computing power and communication capacity will be hundreds of times.

Let's take a look at the challenge of business collaboration: the technology stack of the emerging industry of the Meta Universe is very complex, essentially because it uses all the IT technologies that have matured and are evolving in the past few decades. This also leads to very long business cooperation links in new industries. We have carefully sorted out the industry in the past period of time, and gradually simplified the industry model to five links. At present, most customers and partners are in these five links. A previous guest mentioned how much it costs to be a digital person, virtual space, etc. The biggest problem is that the efficiency of cooperation is extremely low, which makes it costly and the effect is not ideal. How to improve our efficiency and make more useful things at the same time is our urgent business cooperation problem.

After discussing the challenges of business collaboration, let's look at how to solve the technical problems. As for technical issues, there is the challenge of computing power, computing scale and communication. Alibaba Cloud has built the Dragon computing platform for many years. Today we upgrade to the CIPU platform in order to solve the technical challenges of computing performance and communication capability. In addition, in the past few years, we have also increased the number of GPU instances by 100 times to meet the challenges of AI and graphics computing, and the size of GPU computing resources has also exceeded 1700P.

On this page, you can see the improvement of computing and communication capabilities brought by the fourth generation of DPCA computing platform.

This page shows the huge advantages of scale, stability and flexibility brought by upgrading to CIPU.

This page is the seventh generation GPU instance that we released last year. This instance has significantly improved the performance of AI computing and visual computing, and supports the second generation real-time ray tracing engine. It is a very suitable computing instance for the meta universe business.

Although it seems that the technical challenges are still very difficult, we have found the direction, so it is relatively easy to achieve breakthroughs. But the bigger challenge is business collaboration.

How to get through business collaboration? You may think that cloud computing vendors provide computing power. In fact, we think this is far from enough for cloud computing vendors. Why is it far from enough? Because cloud computing has a great advantage, connecting the upstream and downstream of demand can connect all demands here and help everyone do better.

Therefore, in addition to making our IaaS products better, we also hope to make the five cooperation links mentioned in the previous section of the entire industrial chain more efficient, and then have a cooperation system. This collaboration system is mainly divided into three layers: one of them is that we will solve the common problems in the industry at the interface layer, such as improving the rendering efficiency and ability of 3D engines, improving the ability of coding and streaming, etc; Further, we cooperate with partners to build a collaboration platform, and we hope to use this platform to make the complex collaboration mentioned earlier more efficient; Finally, we also build industry editor matrix or industry SaaS matrix for the whole industry, so that more industries can build their own metauniverse applications through low code.

Here, let's briefly introduce two products of the collaboration platform: cloud 3D engine editor platform and cloud XR platform.

About the collaboration platform, many guests mentioned the 3D engine. In fact, the 3D engine plays a very pivotal role in it. All games, applications of the metauniverse, and 3D applications need to be developed with the 3D engine and the editor of the 3D engine. However, there is still a threshold for 3D engines. The threshold is very large in terms of technology, computing power, and the whole business collaboration process. Maybe every such application needs to carry hundreds of gigabytes of material, which is equivalent to a very large amount of computing.

For example, if we make a thing or bake it, it may take dozens of hours, and developers will have to wait for dozens of hours, which is very inefficient. We hope to solve this problem by using the cloud. You can develop business on various types of terminals. The development environment is not limited, and the computing resources are not limited, so the development efficiency is greatly improved.

03 The cloud XR platform is released to support users' efficient deployment and immersive application

Where are the developed products deployed? In fact, we have released a cloud XR platform, which can support people to develop their own applications and put them directly on this platform. It is not necessary to develop applications like before. It is very efficient to adapt hundreds of these terminals. It may take five minutes to build a cloud metauniverse application.

With regard to some features of the cloud XR platform that you may pay more attention to, such as OS type, engine type, and terminal type support, we have also completed the support for nReal glasses. Welcome to try.

You may be concerned about the impact on the development process. Basically, there is no impact. Whether you are developing 2D applications or 3D applications, there is no big difference from the past.

In the end, no matter whether we create and develop on the cloud or deploy on the cloud, everyone's link has not changed. In fact, we also have many immersive applications on the cloud.

The picture above is a metauniverse concert of our customers, and the content of the work is from our partners. This is the effect of some pictures of the Yuanuniverse concert they held on September 4, which is rendered in real time on the cloud.

I talked about the automobile industry earlier. In fact, these immersive applications, or visual computing applications, can actually help many industries solve problems.

For example, one of the hottest industries now is automatic driving. But you will find that there is a huge problem with automatic driving. It requires tens of billions of kilometers of road test, but China's leading enterprises may only have tens of millions of kilometers of road test, and the road test mileage of enterprises that start late is shorter. At the same time, the cost of these road tests is also very high. We need to buy tens of thousands of cars, hire tens of thousands of drivers and security officers to do various road tests, and the amount of data accumulated is also quite limited. But if we simulate it in the virtual world, the effect will be extremely fast. We may say that tens of thousands of GPU instances are driven on the cloud, and tens of thousands of GPUs are used to render such scenes, We use tens of thousands of virtual cars to run, and we run 24 hours a day.

Because the flexibility of the virtual world can be defined, various roads, lighting and environments can be defined, which makes the probability of detecting various bugs in the automatic driving program much higher than under the conditions of the real road test, which can positively improve the production efficiency.

In addition to automatic driving, we also pay attention to the health industry. This is one of our partners. They have become the most advanced medical digital people in the world. Most of their research direction is to study the human body. In fact, many medical schools and hospitals now need such mannequins to do preoperative research planning, learning, etc.

Through visual computing, we can easily deliver this service to every institution or individual in need to help them effectively achieve the purpose of research and learning.

It is believed that it is the vision and goal of every colleague to bring the most real "immersive experience" to everyone with interactive technology and surging computing power, break the two-way challenge of technology+business cooperation, open up the upstream and downstream of the industry, and help it achieve efficiency. We also hope that one day, we can see that interactive technology will truly bring new industrial opportunities to the whole chain of the whole industry.