Dialogue with NodeReal co-founder Ben Zhang: Currently, very few service providers can truly reach the core of "Web3 version of AWS."

Interviewer: flowie, ChainCatcher

Guest: Ben Zhang, Co-founder & COO of NodeReal

During his time at Binance developing BNB Chain, Ben Zhang, Co-founder & COO of NodeReal, realized that with the emergence of many high-performance chains like BNB Chain and Solana, the Web3 industry would increasingly demand high-performance infrastructure, urgently needing infrastructure service providers similar to AWS.

In 2021, Ben Zhang participated in founding NodeReal, which focuses on blockchain infrastructure services, as COO. NodeReal has gathered a group of chain experts and cloud computing specialists from BNB Chain, AntChain, AWS, Microsoft, Alibaba, etc., hoping to create a "Web3 version of AWS" in the future.

However, like NodeReal, there are many blockchain infrastructure service providers aiming to become the "Web3 version of AWS," with well-known projects like Infura and Alchemy aligning with this concept.

What differentiates NodeReal? How are the so-called "Web3 versions of AWS" performing? How far are we from a true "Web3 version of AWS"? Around these questions, ChainCatcher recently interviewed Ben Zhang.

Ben Zhang stated that most of the current "Web3 versions of AWS," represented by Infura and Alchemy, focus on node and API services at the application and access layers. In contrast, NodeReal aims to provide full-stack services that include the underlying, access, and application layers. In addition to node and API services, NodeReal also offers lower-level services such as application chain construction, allowing developers to address all their development needs at once. "As a core developer of BNB Chain and BNB Greenfield, NodeReal has accumulated a wealth of core blockchain technical capabilities."

In Ben Zhang's view, the current "Web3 version of AWS" is still in its early stages, and there is a long way to go to achieve the maturity of AWS. Service providers like Infura and Alchemy, which focus on node and API services, are not genuinely trying to become AWS; only a few companies in the industry, like Eclipse and Kaleido, are seriously working on a Web3 version of AWS.

Currently, Web2 giants like AWS and Alibaba are also expanding into Web3, and in the long run, the Web3 version of AWS will have many competitive and cooperative relationships with them. The key to winning for the "Web3 version of AWS" is not rapid expansion, but finding opportunities for killer applications and solidifying the underlying infrastructure.

Additionally, regarding this year's popular narratives in Web3 such as zk and decentralized storage, Ben Zhang shared his views, and the following is the full dialogue:

1. ChainCatcher: Could you first introduce the history of your team's development and the background of the core team members? How did you personally get involved in Web3?

Ben Zhang: My Chinese name is Zhang Xiaoguang. I was the original head of the BNB Chain ecosystem and am now the Co-founder & COO of NodeReal.

I mainly studied high-performance computing at Shanghai Jiao Tong University. After graduating in 2004, I have been working in high-performance computing, first joining the enterprise database software company Sybase as an engineer. In 2010, Sybase was acquired by SAP, and I then joined Morgan Stanley, where I worked for about eight years.

Through the Chief Technology Officer Summit at Morgan Stanley, I was exposed to many cutting-edge technology startups, especially blockchain companies. At that time, blockchain applications were just starting, and many decentralized applications were emerging with various technical architectures.

Fortunately, in 2019, I had the opportunity to join Binance, initially responsible for Binance Chain (renamed BNB Beacon Chain). In 2020, I developed Binance Smart Chain (renamed BNB Smart Chain) with several colleagues. Later, these two chains merged and were unified under the name BNB Chain. We were lucky to catch the DeFi Summer and the explosion of NFTs, and the overall usage of BNB Chain was quite good.

During this phase, we found that with the emergence of many high-performance chains like BNB Chain, Ethereum, and Solana, there was a significant demand for high-performance infrastructure in the industry. However, the Web3 industry currently lacks mature cloud computing service providers like AWS that can offer developers various types of technical architecture services. So we identified this market pain point and founded NodeReal, hoping to become the AWS of the Web3 space.

Our team background can be roughly divided into three categories: one mainly has a blockchain background, such as BNB Chain and AntChain; another comes from cloud providers like AWS, Microsoft, Tencent Cloud, and Alibaba; and the third consists of members experienced in handling high-concurrency businesses, some of whom come from Morgan Stanley and other companies.

2. ChainCatcher: NodeReal's products are mainly aimed at developers, which may still be relatively unfamiliar to many Web3 users. Can you introduce NodeReal and its products in simpler terms?

Ben Zhang: Let's start from the developer's usage scenario. Many Web2 developers actually don't quite understand the difference between Web3 and Web2. In fact, there are many similarities in the development architecture of Web3 compared to Web2, but it also adds some complexity.

Generally speaking, Web2 development requires deploying a three-layer or four-layer architecture, such as backend, middleware, frontend, etc. Then you need to deploy everything to AWS cloud storage, manage databases, etc., which is quite clear. However, Web3 development becomes more complicated; not only do you need to deploy various functional architectures like in Web2, but you also have to deal with more complex backend storage issues, such as storing some data in databases and some on-chain.

Therefore, whether for Web3 or Web2 developers, we hope to lower the development learning curve while providing easier access to backend and blockchain data. NodeReal has already deployed the underlying parts such as node services and API interfaces, and of course, Infura and Alchemy are also doing this.

Secondly, as large applications or ecosystems, they not only need to access the blockchain but also need to operate their own blockchains and may even require hosting services. We also want to be a provider like AWS, offering Blockchain-as-a-Service.

Overall, we divide our products into three layers.

The first layer is the underlying Application-Specific Chain/L2-as-a-Service and Explorer-as-a-Service. Why do we have this capability? As core contributors to BNB Chain, we have accumulated solid technical capabilities in this area.

The second layer is the access layer products, which platforms like Alchemy are working on; the third layer is tooling APIs, providing Indexing APIs, Enhanced APIs, etc. In summary, the higher the layer, the closer it is to developers, while the lower layers are more core-focused.

3. ChainCatcher: What differentiates NodeReal from Infura and Alchemy? As a latecomer, where are NodeReal's opportunities? What is the key to winning?

Ben Zhang: The entire tech stack is quite different. Infura and Alchemy focus more on the application layer, while NodeReal aims to provide developers with one-stop technical services that include both the underlying and application layers.

From a blockchain technology perspective, the lowest layer is the blockchain kernel, followed by the blockchain access layer, and then the blockchain application and developer layer. Infura and Alchemy are essentially working on the application and access layers without providing Blockchain-as-a-Service.

But NodeReal not only works on the application and access layers but also on the more core underlying layers. This positioning relates to our team background, as many team members have previously worked on hardcore blockchain projects. Additionally, from the perspective of user understanding, we hope that users adopting the underlying services won't have to struggle to find third parties for the access and application layers.

Looking back, this is very similar to the early development of AWS. AWS initially provided Virtual Machines and later offered many IaaS (Infrastructure as a Service), PaaS (Platform as a Service), and SaaS (Software as a Service) services, ultimately meeting various needs of developers.

Currently, most developer service providers are actually offering PaaS or SaaS services. However, we still hope to eventually cover developers' needs across all layers, similar to AWS and Alibaba Cloud.

What is the key to winning? From the perspective of internet development, there were many bubbles before 2000, where anything related to the internet was considered innovative, but eventually, the bubbles burst. It wasn't until after 2000, with the emergence of cloud computing and the rise of social applications, that the internet truly unleashed its power. Web3 is similar; we still don't know what the core application scenarios will be in the future. Therefore, the key to winning is not rapid expansion but finding opportunities for killer applications and deeply collaborating with killer application partners to solidify the underlying infrastructure for sustainability.

4. ChainCatcher: Recently, the BNB Chain released the white paper for the decentralized storage network BNB Greenfield, which has attracted much attention. As one of the core developers of BNB Chain and Greenfield, what do you think is the biggest innovation BNB Greenfield brings to decentralized storage networks?

Ben Zhang: Our core contribution to Greenfield lies in data, such as on-chain indexing and data access.

Before understanding BNB Greenfield, we need to know what BNB Chain's starting point for BNB Greenfield is.

BNB Chain's approach is different from other public chains. BNB Chain continuously expands its block size, and it currently may have the world's largest block, with a very large amount of data that can be stored in each block.

We conducted a data report showing that in 2022, BSC generated over 10 million blocks, causing the historical data (blocks, transactions, receipts, and code) storage size to increase from 658GB to 1.07TB (a 63% increase). The total storage size of the entire MPT trie tree also increased from about 205GB to about 361GB (an increase of about 76%). The total storage size of the pruned BSC full node reached 1.60TB, nearly 2.46 times that of Ethereum (the pruned Ethereum full node occupies 650GB).

Such a massive amount of data is challenging to store, so the team began looking for better storage solutions, naturally leading to the idea of decentralized storage solutions. Currently, there are decentralized storage solutions like Filecoin, Arweave, and Storj, each with its advantages and disadvantages. However, for BNB Chain, directly adopting existing solutions on a large scale poses some challenges.

First is the performance challenge. BNB Chain has a large user base, recording about 1 million daily active addresses in 2022. At this point, the performance requirements for the data access layer are very high, requiring millisecond-level speeds. The retrieval process of Filecoin is not always as fast as IPFS fixed services or traditional cloud storage; the minimum file size accepted by its storage providers is usually several GB. The Arweave network also faces bottlenecks in effectively uploading and retrieving large files. If you use Storj services and try to upload an entire folder of photos, you will find the response time slow and the waiting time long.

Second is the inability to integrate seamlessly. Filecoin, Arweave, and Storj require BNB Chain users and dApps to have FIL, AR, or STORJ to use their storage services, leading to challenges with inconsistent address patterns. BNB Chain wants a solution that allows all users to access all storage services using the same BSC address, thereby reducing barriers to large-scale adoption.

Third is the inability to properly manage privacy and access control. Files stored on Filecoin and Arweave are publicly accessible, meaning anyone can retrieve those files.

Without a good solution, BNB Chain hopes to integrate innovation to provide one-stop services for applications in its on-chain ecosystem, such as games and the metaverse, rather than searching for various infrastructures everywhere. Importantly, it ultimately aims to empower its token BNB.

5. ChainCatcher: From the perspective of developers, what do you think the short-term and medium-to-long-term market landscape for rollup solutions like Optimistic Rollup, zk rollup, and zkEVM will look like? How should developers choose a rollup?

Ben Zhang: These solutions certainly have their pros and cons, and it's hard to say which is better. Our view is that at this stage, we are more optimistic about Optimistic Rollup, as zk and zkEVM have a high technical threshold and are far from high-performance rollups and industrial applications.

Some may question the security of Optimistic Rollup, thinking it is off-chain, but off-chain does not mean it is insecure, as it has more fault proofs to ensure gradual backups, and performance can be very stable. Therefore, this is also our preferred rollup service solution.

For developers, when choosing a rollup, they should consider where there is better support, a better ecosystem, and developer tools, etc.

6. ChainCatcher: Currently, NodeReal is mainly developing on BNB Chain and Aptos. What are the main considerations behind this?

Ben Zhang: On the surface, it seems that way, but internally, we actually divide the ecosystem into two major categories.

One category is truly large-scale ecosystems, such as BNB Chain and the Ethereum ecosystem, which includes Polygon, Arbitrum, Optimism, etc.

The other category consists of emerging layer 1s, especially those focused on high performance and aiming for large-scale adoption. The reason we initially chose Aptos, besides being familiar with their team, is that our early development philosophies align; we both want to focus on high performance.

So why does it seem like we are mainly developing on BNB Chain and Aptos? Because we hope to collaborate deeply with their ecosystems to provide one-stop services. However, this is the current situation; we plan to delve into the Ethereum ecosystem next, and we have just released an application chain based on the Optimism tech stack.

7. ChainCatcher: NodeReal expanded its application chain services last year. What common pain points exist in application chain construction and usage? How does NodeReal help users address these pain points?

Ben Zhang: Many people think that a Web3 application chain is just an application chain like Axie Infinity, but in fact, application chains have expanded. There are generally two typical types of application chains. One is completely independently operated, like Axie Infinity, which can be attached to an L1 main chain or not. The other type, like rollups, needs to be attached to an L1 main chain.

So what are the pain points for project parties in building application chains?

First is the lack of chain experts. Projects that attempt to build application chains on their own can easily run into problems. Axie Infinity took a long time to develop, but encountered significant issues throughout the process, and the resulting application chain faced various security incidents.

Second is the lack of business experts. The business is very complex; how to ensure the chain effectively supports the business requires not only chain experts but also business experts. We need to communicate clearly with clients about their needs, why they want to put these businesses on the application chain, and what specific requirements are needed, such as how to configure gas fees, etc. At this point, some business experts are needed who understand business logic and the characteristics of the chain to effectively implement business logic on the chain, which is very complex.

Third is not knowing how to operate the chain. Chains require operation, but many project parties lack experience and do not know how to expand their ecosystems.

Overall, project parties need a more comprehensive and simpler one-stop service, which is why we aim to provide full-stack services. We hope project parties can resolve everything in one go on our platform, rather than having to find different suppliers for cement, bricks, etc. This is precisely the capability that Alchemy and Infura do not possess.

In terms of application chain development services, we expect to explore 10-20 deeply cooperative projects this year, addressing the most common and core pain points one by one, solidifying the infrastructure, and then serving a large number of users.

8. ChainCatcher: NodeReal has claimed to be the fastest node service provider on BNB Chain and Ethereum. How should enterprise clients choose service providers? What are NodeReal's strengths and weaknesses?

Ben Zhang: For our enterprise clients, those with a large user base should follow the three "S" when choosing RPC services.

First is Speed. For clients with a very large user base, speed significantly affects the experience, and we can achieve speeds of tens of milliseconds.

Second is Scalability. Many users currently do not pay attention to scalability because many chains have a small user base, so there is no need to compete. However, when the user base grows, it can easily impact performance. The network must be designed to support a large number of transactions per second (QPS) without sacrificing its effectiveness or security.

Third is Stability. Poor stability levels in dApps can lead to inconsistent user queries, but response times from different regions should always be stable and predictable.

Unlike serving ordinary developers, we place greater emphasis on helping enterprise clients filter their needs before providing node services. We work with them to fill out a complex questionnaire to understand their business and real needs.

9. ChainCatcher: The centralization issue of node service providers poses significant risks to users and platforms. The Infura outage caused widespread paralysis of the Ethereum network, and many node service providers, led by Infura, are expanding into decentralized protocols. What are your thoughts on this? What is NodeReal's current architecture and future plans?

Ben Zhang: In the long run, there indeed needs to be more decentralized protocols, and we are also applying to join decentralized protocols to explore future developments. However, balancing decentralization, performance, and security is very challenging; generally, achieving decentralization results in poor performance.

Therefore, in the short term, we still consider the high availability and risk resistance of individual node service providers, as well as the need for large networks like Ethereum to use enough node service providers to achieve artificial decentralization. Even if one service provider in the network fails, there are others available to prevent network paralysis.

10. ChainCatcher: There are many infrastructure service providers aiming to become the "Web3 version of AWS." How are they currently performing? How far are we from a true "Web3 version of AWS"?

Ben Zhang: To determine whether someone is genuinely working on a Web3 version of AWS, we can evaluate them based on the core principles of AWS.

AWS solved several key problems, first addressing computing and storage, allowing developers to avoid deploying hosts. The computing and storage of the internet correspond to the EVM and data of the blockchain, but very few companies are working on this core aspect.

Secondly, AWS built many higher-level developer services based on the underlying computing and storage, such as APIs and API marketplaces. This allows both large and small developers to deploy on AWS in a one-stop manner, making the one-stop aspect crucial.

Earlier, I mentioned the three layers of the development architecture: the underlying layer is the blockchain, the middle is the access layer, and the upper layer is the application layer. Currently, many people are working on the application layer and are not addressing the lower layers like AWS does.

So how far are we from a Web3 version of AWS? I believe we are still quite far. Companies like AWS and Google not only form a heavy technical architecture service but also provide storage services, security management services, etc., which involve very complex content.

Currently, we believe that only a few companies like Eclipse and Kaleido are seriously working on a Web3 version of AWS. Companies like Infura and Alchemy are essentially just providing node and API services and are not genuinely trying to become AWS.

Moreover, the future form of Web3 is richer, and creating a Web3 version of AWS will be more challenging. However, it is essential to focus on specific scenarios and gradually expand.

11. ChainCatcher: AWS is also expanding into Web3. What will the competitive landscape look like between AWS and those aiming to create a Web3 version of AWS in the future?

Ben Zhang: In the long run, it will be a competitive and cooperative relationship, but currently, it is more of a cooperative relationship. AWS and Alibaba are our partners, and we recently collaborated with AWS to host the Hong Kong Innovation Meetup event.

From my perspective, AWS or Alibaba's positioning will be more foundational, as they hope to have more users utilizing their computing, storage, and even hosting services. They are unlikely to have the bandwidth to cover all the underlying technical services of Web3, such as running node services and chain data services. Additionally, the upper layers of Web3 are very diverse, and AWS cannot do everything and won't be able to. Therefore, we may focus more on the underlying technologies and application scenarios of Web3.

Furthermore, AWS and Alibaba have many Web2 users and scenarios that can complement us, and helping Web2 clients better utilize blockchain is where we excel.

12. ChainCatcher: NodeReal has always emphasized uncovering users' true needs. In the face of developers' demands, can you share some pseudo-needs you have identified?

Ben Zhang: That's a good question. First, we are very strict in demand management and do not spend time creating products that do not yield benefits or feedback. However, we have identified some pseudo-needs in serving users, such as the belief that supporting more chains is better, which is a pseudo-need, as having multiple chains does not necessarily bring in more new users.

13. ChainCatcher: In serving a large number of developer needs, have you glimpsed any innovative trends in the Web3 industry?

Ben Zhang: First, everyone is considering whether there is high-performance infrastructure to meet the next wave of user explosions. Currently, many L1 chains may have fewer than 5 million daily active users. Are we prepared for when daily active users reach 10 million or even 100 million?

The second is where the original applications are. Previously, so-called innovations involved switching to different chains to make simple changes, but now we need truly original algorithms. For example, in DeFi, we may need to introduce new, more efficient computations.

The third is how Web2 expands into Web3, transitioning from point attempts to platform attempts. Some Web2 giants are still in the early stages of trying out single points without turning them into a platform.

14. ChainCatcher: As a Chinese Web3 entrepreneur, what experiences and advice can you share with other Chinese Web3 entrepreneurs during this cycle?

Ben Zhang: Web3 and Web2 entrepreneurship share commonalities and differences.

The commonality is that both require innovation. When the internet truly exploded, China had many original technological innovations. However, currently, Chinese Web3 entrepreneurs seem a bit too eager and have not taken the time to think about how to achieve original breakthroughs. In fact, there are still many opportunities for innovation in Web3 now and in the future.

At the same time, it is essential to consider what your "beachhead" is, in what kind of "beachhead" you can play a significant role, and to firmly establish it. When BNB Chain rose, it captured two crucial "beachheads": DeFi and gaming. We identified these two application scenarios and truly established ourselves in them.

The difference between Web3 and Web2 is that Web3 can take off faster because it is easier to find co-creation community partners who support each other, rather than having to go it alone as in the early stages of Web2.

Chinese Web3 entrepreneurs need to continually ask themselves if they have enough passion and patience to persist for 5-10 years or more, waiting for the real explosion to come.