Abstract
DeepBrain Chain, a distributed high-performance GPU computing power network, may deeply empower the metaverse.
Introduction
Neal Stephenson’s novel “Snow Crash” in 1992 depicts an independently- existing virtual world that is parallel to the real world. In this fictional world, people can live as they were in reality. Coincidentally, Chinese scientist Qian Xuesen also imagined the world and named it the “spiritual realm.”
Nowadays, the “metaverse” has become the most popular topic.
At present, the metaverse an “advanced” sector at the forefront of technology. It profoundly integrates virtual reality, 5G, AI, cloud computing power, big data, blockchain, and other technologies, which together form a field with various disciplines and technologies intertwined.
Some people use Karl Marx’s theory in “Wage Labor and Capital” to analyze the metaverse: elements like 5G, AI, computing power as productivity; nature data, human data, IoT data, and digital raw data as means of production; with the blockchain playing a role as production relationship.
The productivity sector is the fundamental element for the healthy operation of the metaverse. In this sector, computing-based 5G, AI, and other factors need computing power as the primary resource support. Then, in the face of the complex ecosystem in the metaverse, who can satisfy its large computing power gap and provide in-depth empowerment?
1. Computing power in the metaverse era
If we regard the metaverse as a car, energy (electricity or fuel) is the fundamental resource to guarantee its operation. While in the metaverse world, computing power plays the role of energy supply. Only a steady stream of computing power supply can ensure the operation of the metaverse.
The metaverse is a new type of Internet application and social form that integrates network communication, extended reality, digital twin, blockchain, AI, and other new technologies. The ecology in the metaverse already has a complete series of economic logic, data, content, and IP. It is a real-time digital world that is constantly online, keeping refreshing, and allowing every user to produce and edit content, which means that the meta-universe is a car that keeps motoring “eternally.”
The “meta-universation” of the physical world means infinitely rich production factors, infinitely sufficient productivity, and highly improved social production efficiency of human collaboration, as well as a near-limit expansion of human online users capacity. With an almost unlimited supply of a sea of content, AI must serve as the primary automated content creation method. As AI is needed to run through the ecological operation of the metaverse, it will lead to an exponential increase in the consumption of computing power in the metaverse. Also, keeping online and content creating for large-scale users requires massive computing power as support, which means that metaverse has almost infinite computing power requirements.
The current computing power system is still a Web 2.0, centralized system. The mainstream cloud computing service providers usually concentrate computing power in multiple data centers composed of hundreds of thousands of servers with CPU as the core. In this way, it continuously provides computing services for the global network. With the surge of the market demand, the cloud computing service providers will also further develop the hardware. However, the overall price level of computing power is still relatively expensive in a price market under suppliers’ control. For example, a single model training for some AI projects will cost hundreds of thousands of s dollars. Similarly, the amount of computing power produced by this method is difficult to form a qualitative change in a short period. Meanwhile, because of the hardware expansion problems, the amount of computing power resources they provide is quite limited.
Usually, the computing power providers may build computer rooms in remote areas due to training model factors like cost, natural disasters. That means there may be a degree of delay in obtaining computing power for some computing power users, which is also difficult to satisfy some application scenarios with higher computing requirements. The centralized computing power supply system has a low fault tolerance rate. Once malfunction occurs, it may cause a considerable impact on sectors that rely overmuch on computing power. We can see that the current centralized computing power system is far behind the required growth rate of data and algorithms in the metaverse ecology.
With the further development of blockchain technology, the distributed collaboration system based on blockchain is expected to reshape the traditional centralized resource supply and distribution system, and fundamentally solve the problems of trust and efficiency. The distributed system based on the blockchain will change from the conventional single-point way to the multi-point distributed supply, where anyone can become the supplier and resource acquirer of the supply system, with no limits under geographical, national, and other factors.
The distributed supply system can conduct unlimited supply expansion, without being affected by single points of failure, which has a higher fault tolerance rate and lower price. In the future, as 5G+AI, edge computing, and other technologies play an increasing role in the metaverse, only a distributed computing power supply system can meet the supply requirements for computing power with high performance, high density, low latency, low cost, and full coverage. Meanwhile, the distributed computing power supply system will rely on a trillion-level blue ocean market.
2. Who is the new infrastructure in the metaverse?
Among the existing infrastructure sectors in the distributed computing power field, people place high hopes on DeepBrain Chain, RNDR, and Dfinity. Who is the new infrastructure for computing power in the metaverse era? The part below will present an analysis of the three potential ecosystems one by one.
Both DeepBrain Chain and Dfinity are decentralized computing platforms, which means they are able to, and both are in a position to provide computing power services for the demanders. They both are essentially blockchain systems based on computing power, building decentralized cloud computing platforms through blockchain technology and connecting to many nodes distributed around the world to construct a distributed computing power hub. After using idle computing resources, the cost of computing power has dropped significantly, and the overall computing power supply is highly efficient with higher fault tolerance.
Besides the disparity in the structure of the chains and the operating methods of every role in the ecology, an essential difference between DeepBrain Chain and Dfinity is the hardware source of computing power. In the Dfinity system, the computing services provided are mainly from the CPU computing power. In this way, Dfinity can be regarded as a decentralized CPU computing power market, while the computing resources provided by DeepBrain Chain mainly come from the GPU.
CPU is presently the main hardware foundation for mainstream centralized cloud computing service providers to produce computing power. Although both CPU and GPU can output computing power, CPU plays more role in complex logic calculations, and GPU can perform large-scale parallel calculations as a processor with hundreds or thousands of cores, which is more suitable for visual rendering and deep learning algorithms. The calculation method provided by GPU is relatively faster and cheaper than CPU, with only one-tenth of CPU’s computing power cost usually.
Currently, GPU computing power has deeply integrated itself into artificial intelligence, cloud games, autonomous driving, weather forecasting, cosmic observation, and other high-end scenarios which require a large amount of calculation. Especially with the concentrated eruption of such high-end industries, the market’s demand for GPU computing power will be much higher than CPU in the future.
Therefore, DeepBrain Chain is a distributed heterogeneous high-performance computing power network that gathers a large amount of GPU computing power, which can further reduce the cost of the computing power demand of the high-end technology industry and satisfy their requirements for computing power quality. For instance, AI model training can further reduce costs through DeepBrain Chain’s cheap, high-performance GPU computing power.
On the other hand, Dfinity’s CPU computing power network aims at realizing blockchainization of popular network application requirements, such as decentralizing information websites and chat software, which is also necessary to build the Web 3.0 world.
With Otoy, one of the authoritative companies in the global graphics field, as its constructor, The RNDR network is mainly positioned for distributed rendering. RNDR has built a blockchain-based distributed GPU computing, rendering network, and 3D content market, and it aims to reduce the cost of users with rendering needs by gathering idle GPU computing power. The RNDR rendering network introduces distributed GPU rendering to the 3D graphics industry and builds a foundation for small and even individual rendering users’ utilization.
Hence for the RNDR network, the application of the GPU computing power system is more directional, rather than supplying computing power to all scientific and technological fields. DeepBrain Chain is more suitable for various technology fields to obtain high-performance and cheap GPU computing power.
From the perspective of building a metaverse, graphics card manufacturers represented by Nvidia believe that GPU determines the actual effect that meta-universe as a virtual world can bring to end consumers to a large extent. With the expansion of the user capacity of the metaverse and the increase of the large-scale users’ demand for keeping online and creating led by AI, the rigid demand for high-performance computing power will be endless. DeepBrain Chain is hopeful to further satisfy the computing power gap in the operation and development of the meta-universe world through a distributed high-performance GPU network, which will be the future opportunities for DeepBrain Chain.
In conclusion, DeepBrain Chain performs better to meet the rigid demand of computing power in the metaverse ecology than Dfinity and RNDR network and is hopefully becoming the major computing power infrastructure in the metaverse era.
3. How will DeepBrain Chain empower the metaverse era?
With a GPU computing power supply capability, any role can become a computing power node of the DeepBrain Chain. It means that DeepBrain Chain will allow GPU computing power servers worldwide to become its nodes, and finally connect them into an open and continuously expanding platform based on incentives. Additionally, DeepBrain Chain shows potential to draw all high-quality computing power providers in series, which makes itself supported by vigoroso computing power resources, and proves to be the world’s largest GPU distributed computing network.
Also, DeepBrain Chain itself is a decentralized computing power network maintained by nodes from around the world, which means that it can expand computing power business worldwide, and allow users from any region and country to obtain computing power through it. Therefore, DeepBrain Chain possesses global service capabilities and powerful computing resources, which will be the foundation for DeepBrain Chain to profoundly empower the metaverse and AI era.
Another advantage of DeepBrain Chain is that the global computing power network can automatically transform into metro and edge nodes that meet the nearby computing needs, filling the gaps of traditional cloud computing servers. Therefore, DeepBrain Chain can provide a real-time computing power supply for various scenarios. As the outline of the metaverse ecology becomes clear, the metaverse will continue to undertake the functions of the physical world, and further enhance the human’s efficiency of collaboration and production by online methods. It is foreseeable that in addition to the increasing demand for computing power in the metaverse ecology, a large number of real-time AI interaction scenarios also have higher requirements for real-time computing power, and DeepBrain Chain will be the essential empowerment foundation in the operation of the meta-universe ecosystem.
At present, DeepBrain Chain has provided global computing power services for nearly 50 universities, more than 100 technology companies, and tens of thousands of AI developer groups, with dozens of AI developers (enterprises) deploying high-performance GPU cloud platforms based on the DeepBrain Chain network. In the next few years, DeepBrain Chain is committed to deploying more GPU computing nodes around the world on the supply side; supporting more companies or developers to deploy AI cloud platforms based on DeepBrain Chain on the demand side; together building the world’s most important distributed high-performance computing power network in the “metaverse+AI era”.