Dialogue with the Conflux Team: Technical Path, Public Chain Competition, and Industry Cycles

Author: Gong Quanyu

With the continuous explosion of public chains, more and more solutions aimed at improving transaction efficiency have emerged, enhancing the performance of public chains from different levels such as sharding and sidechains.

Recently, Conflux adopted a tree-graph structure to enhance the transaction processing capacity of its main chain, with the testnet TPS exceeding 3000, and the mainnet expected to go live by the end of the year. It is worth mentioning that Turing Award winner Academician Yao Qizhi is the Chief Scientist of Conflux, and many team members come from top universities such as Tsinghua University and the University of Toronto.

As a result, Chain Catcher recently interviewed Dr. Wu Ming, CTO of Conflux, and Zhang Yuanjie, COO, discussing Conflux's technical characteristics, development plans, and industry perspectives, hoping to provide readers with insights and inspiration for understanding public chains.

1. Tree-Graph and Technical Path

Chain Catcher: Public chains are the underlying infrastructure of the blockchain industry and are also the most competitive field. What are the main factors currently restricting the development of public chains? Where do you see opportunities?

Conflux: We have found that the throughput of public chains has become a significant obstacle to the development of the blockchain industry. Many decentralized applications with practical commercial value are limited by the performance constraints of public chains like Ethereum, making it difficult for them to be effectively implemented on public chains.

Currently, although many public chains have proposed solutions for transaction efficiency, they generally sacrifice security. Indeed, there is a ceiling on efficiency under the premise of ensuring decentralization, which is limited by network synchronization bottlenecks, but current technology has not yet touched this ceiling. We believe that tree-graph technology can break through the current bottleneck.

In addition, there are a series of issues regarding the security of smart contracts and virtual machines in the blockchain, as well as user privacy protection, which are all opportunities we see.

Chain Catcher: You recently changed your technology from DAG to tree-graph. Could you briefly introduce the characteristics and advantages of tree-graph technology? What are the differences between tree-graph and DAG?

Conflux: In a DAG structure, each node has only one type of edge, while in Conflux's tree-graph structure, each node has two types of edges: parent edges pointing to the parent node, and there can only be one parent; reference edges can point to other nodes, expressing the Happens-Before relationship between different nodes. If we only look at the parent edges, the structure of the ledger is a tree; if we consider both parent edges and reference edges, the structure of the ledger is a graph. The tree-graph structure refers to a structure that contains a tree within a graph.

Note: Solid arrows point to parent blocks, dashed arrows point to reference blocks

The tree structure formed by parent edges is very important for Conflux's consensus mechanism. It allows machine nodes in the Conflux network to reach consensus on the total order of blocks and transactions in a decentralized and secure manner. The structure of the tree determines the selection of the main chain, and the selection of the main chain determines the total order of transactions. Therefore, we believe that referring to Conflux's ledger structure as a tree-graph structure better reflects the technical essence of Conflux.

In contrast, some existing DAG projects, such as IOTA and Byteball, either do not support the total order of transactions, thus failing to support smart contracts, or cannot achieve consensus on the total order of transactions in a decentralized manner, sacrificing security.

Chain Catcher: There are many solutions for improving transaction efficiency, such as sharding, Plasma, and sidechains. How do you evaluate the prospects of these solutions? What will the market landscape look like in the future with scalable consensus algorithms based on tree-graph structures?

Conflux: These solutions do not conflict with Conflux's technical solution; rather, they are complementary. Conflux is a Layer 1 technical solution, while sidechain technology belongs to Layer 2 solutions. Sidechain technology itself needs to rely on Layer 1 public chain technology. Moreover, Conflux is designed to be perfectly compatible with Layer 2 sidechain technology.

Although sharding technology belongs to Layer 1, such technology can also be applied to Conflux. Conflux addresses how to improve throughput without sharding, and we can certainly further enhance throughput using sharding.

However, most current sharding technology solutions sacrifice decentralization and security to some extent. Since transactions within a shard do not interact with other shards, these shards effectively form independent security entities, and their security will be determined by the computing power within the shard, which is significantly less than that of the entire network. Currently, we are also actively seeking a more perfect sharding solution.

Whether it is sidechains or sharding, there is indeed a significant demand for technical solutions like Conflux. Sidechains often require complex interactions with the main chain when users create and exit, which places certain performance demands on the main chain. A main chain capable of supporting a large sidechain ecosystem must provide high throughput.

Additionally, in a sharded system, we cannot simply assume that transactions occur solely within shards. In practical applications, cross-shard transactions are also prevalent. In this case, Conflux's solution will help sharded systems better handle cross-shard transactions, preventing them from becoming a performance bottleneck for the system.

Chain Catcher: What factors do you consider when designing Conflux's token economy?

Conflux: The specific economic model has not yet been finalized, but one point we can share is that we will make significant adjustments to transaction fee distribution. Since Conflux uses a tree-graph structure, the same transaction may reach different blocks, which can actually harm throughput. Therefore, we will avoid malicious competition through our incentive mechanisms.

In some public chains, participants sometimes compete for transactions with high fees, which can lead to a loss in throughput. We will establish a rule based on game theory, allowing miners to maximize their profits by following this rule when packaging transactions, thereby also enhancing Conflux's transaction throughput.

2. Industry Competition and Future

Chain Catcher: Last year saw a large number of public chains go live, and this year, heavyweight players like Conflux are entering the market. The competitive landscape of public chains is becoming increasingly complex. How many public chains do you think truly have market competitiveness? What will be the long-term landscape and evolution of public chains?

Conflux: The status of Bitcoin is undeniable, and Ethereum is also a strong competitor with a very solid community, but its main application scenario, ICO, has been proven to be a false demand, so we should put a half-question mark on Ethereum. As for other public chains that have already gone live, we are not very optimistic and would put a question mark on them.

There are still many academic teams developing public chains worth looking forward to, such as Algorand by Turing Award winner Silvio Micali, Thunder by Cornell University professor Elain Shi, and Dfinity. Additionally, there is the TON public chain developed by Telegram, which has a significant user base and community foundation.

If we view blockchain networks as distributed systems that need to support various applications like ordinary operating systems, it will inevitably lead to only a few network systems being able to support a sufficient number of applications. Only with enough applications can a more stable and powerful network effect be formed. Therefore, we believe that ultimately there may only be one or a few public chains, and whether Conflux will be one of them remains to be seen.

Chain Catcher: Along with the birth of numerous public chains, a large number of DApps have also emerged. Major public chains are doing their utmost to promote their DApp ecosystems, but in reality, most DApps lack real users, and many of the more active DApps are in gaming and gambling. How do you view the current application landing situation?

Conflux: If users wish to achieve their goals quickly and prioritize experience, centralized solutions are not out of the question. However, if this behavior involves significant economic value and potential conflicts of interest due to mutual distrust, decentralized applications become a very good solution.

Currently, the directions for the application of blockchain technology mainly fall into three categories:

1. Payment tools: Due to the high intermediary fees charged by traditional financial institutions, the activity of Bitcoin as a payment tool has been steadily increasing, especially in developing countries where inflation is evident and the banking system is relatively weak.

2. Distributed finance: There are many collateral lending applications and decentralized exchanges on public chains like Ethereum.

3. Gambling: Players engage in mutual competition, primarily occurring on EOS and Tron, due to their fast transaction confirmation times after adopting the DPoS mechanism.

In our view, the technical path will determine the development of downstream ecosystems, such as the number of nodes and the degree of decentralization. Although EOS appears to have high transaction efficiency, if the security of the public chain is not trustworthy, genuine decentralized financial applications will not develop on it.

Currently, there are still relatively few applications in the payment and finance sectors, but I am very optimistic about their potential. Meanwhile, gambling applications are gradually declining as the house takes its profits.

As the performance bottlenecks of blockchain are broken in the future and developers' imaginations are gradually stimulated, the DApp ecosystem will flourish, just as no one anticipated the emergence of applications like Douyin and Kuaishou during the 3G phase of mobile internet. Only when we reach the 4G stage will a sufficiently rich application ecosystem be stimulated.

Chain Catcher: If most solutions succeed in the future and the operational efficiency of mainstream public chains generally improves, where will the competition focus of public chains lie next?

Conflux: First of all, I think this assumption may be overly optimistic. So far, I have not seen any public chain publicly state that it can achieve TPS of 500 or 700 under completely decentralized conditions. However, Conflux's recent stress test reached this level, so I find this assumption difficult to realize.

If this situation does occur, I believe the focus of competition will be on the community. The strength of the community will determine the potential size of the public chain's ecosystem. An important factor in managing the community is the team's original intention. A team with a good original intention will encourage participation in community building and benefit the community ecosystem, rather than incentivizing people to buy tokens or play Ponzi games. If the team's original intention is poor, the community will always be a mob, and once people make money, they will scatter to the next project.

Chain Catcher: What will your community-building path look like moving forward?

Conflux: Currently, the team is primarily doing it themselves because the early stages of the project require a strong core team to launch the underlying architecture, and then gradually promote community building. If an exceptionally talented team does not build the underlying architecture, it will be very difficult for the public chain to succeed. For example, although Ethereum has proposed many beautiful technical visions, it has also faced delays.

The most concrete manifestation of community building is open-source code, placing oneself under the supervision of all developers and users with an open and transparent mindset, and encouraging developers to contribute code. Conflux will gradually open source its code in the future.

Currently, Conflux's user community has adopted a "proof of work mechanism" to incentivize community users and external organizations to contribute. After joining the community, they work for a period of time, then submit a work report to the foundation and community, and all community participants vote together to decide how much reward the contributor should receive. Although this mechanism cannot guarantee complete fairness, it can achieve relative fairness.

Chain Catcher: How many people are currently on the Conflux team? There are voices in the market saying that some members of Conflux are also part-time academic researchers, such as Dr. Long Fan, the CEO, who still serves as an assistant professor at the University of Toronto. Does this affect the operation of the project?

Conflux: Currently, the core public chain team has about 20 members, with over 20 in market and other functional departments. Considering community building, the Conflux team plans to keep the number under 100 in the future.

It is true that several core team members are also engaged in academic research. Blockchain is a very cutting-edge technology that requires a combination of industry, academia, and research. Having a few people in academic positions actually benefits the project significantly.

Chain Catcher: Academician Yao Qizhi is the Chief Scientist of Conflux. What assistance has he provided to the project so far?

Conflux: First of all, Academician Yao has regular communication sessions with the technical team every month or every two weeks, providing guidance on theoretical proofs, especially in areas of game theory and cryptography that he is particularly interested in.

Secondly, Academician Yao is also a board member of the foundation and participates in its governance. Many core team members are his students, so they will not pursue directions that do not align with his expectations.

Finally, as the Chief Scientist, Academician Yao brings a halo effect and convenience to our interactions with external enterprises and government departments.

Chain Catcher: How do you view the changes in the development cycle of the blockchain industry?

Conflux: We believe that the blockchain industry is iterating through three cycles, which may be separate or overlapping.

The first cycle is the technology cycle, where breakthroughs in underlying technology are the main task. Today, the top figures in cryptography, distributed systems, and game theory are participating in this industry, such as Turing Award winner Silvio Micali and alumni from Yao's class. This year, many of the projects mentioned earlier are at the stage of presenting their work, and it's time to see what is real and what is not.

The second cycle is the application cycle. After the ICOs on Ethereum were discredited, the industry is now evolving towards decentralized finance. The gambling applications on EOS also gained some traction for a while, but their momentum is actually quite weak. Therefore, the application cycle will continue to iterate forward as everyone searches for various directions.

The final cycle is the liquidity cycle, which is similar to the cycles of currency development and stock markets. For example, next year, Bitcoin's block reward will be halved, and many people believe this will lead to a price increase.

If we observe the changes in the blockchain industry through these three cycles, I believe the perspective will be clearer and more accurate.