Spot Light | Conflux founder, Dr. Fan Long: we are the most reliable blockchain team

12 June, 2019

It is extremely rare to find a second blockchain startup team like Conflux. Here is why.

Prof. Andrew Yao, the only Chinese Turing Award winner, the founder of the famous “Yao Class” and the dean of the Institute of Interdisciplinary Information at Tsinghua University, is the chief scientist; Last year at the time when the blockchain and cryptocurrency markets were more depressed than ever, this team announced a $35 million funding from Sequoia and the like; and the core members are from the top academic circles, including Tsinghua Yao Class and others, and the global competition winner circle. As Conflux founder Fan Long puts it, “We have a dozen of gold medals here”.

Why these all-stars choose to join the blockchain space? What breakthroughs have they made since last May? How do they see the current cryptocurrency market, the Impossible Triangle bottleneck or the dispute of PoW and PoS? How Conflux will stand out from the competition among public chains?

Recently, 31QU exclusive interviewed Conflux founder Dr. Fan Long and hereby to share some of his insights.

written by /31QU

The Most Reliable Team in the Blockchain Space

From a cryptography lecture at Tsinghua University to publish the paper titled "Scaling Nakamoto Consensus to Thousands of Transactions Per Second", this is the "original story" that has been repeatedly reported.

“We thought we were just preparing a paper.” Fan recalled. After the paper was published last May, his focus was still on research only. But many investor friends advised, don't stop at the paper, do it seriously.

Later on, a conversation with Prof. Yao truly encouraged him to actualize Conflux. “Prof. Yao is particularly supportive.” Fan said, especially with the system theory, “Only the actual adoption and application could grant significance.”

Therefore, the project was launched last June; in December, during a depressed bear market, the news of "Conflux received $35 million funding" was spread out.

At this point, people who were concerned about the blockchain and cryptocurrency industry were turning their attention to Conflux.

Fan first learned about Bitcoin was from a class at the Massachusetts Institute of Technology (MIT), “Bitcoin was a case study.” Fan recalled, during which they discussed how Bitcoin could get through the consensus as well as the advantages and disadvantages of its system design, “to study Bitcoin from a purely technical perspective”.

At that time, Bitcoin stepped out of the geek circle and began to engage with the real world, and the early fans were trying to promote, use and trade it.

"I was very interested in Bitcoin. When it was $5 each, I bought 200 in Mt.Gox, and then I happily sold them when the price hit $50." Fan couldn't help laughing about this first interaction.

"It was intriguing and fun", which was the initial impression of Bitcoin left to 22-year-old Fan.

In addition, he sensed that the Bitcoin technology could potentially be revolutionary.

"Internet enables easier communication, but when it comes to collaboration, especially among strangers, a consensus is also crucial. In short, consensus is a higher level of demand above communication and information transmission." Bitcoin is the first public chain that establishes a consensus, it proved to everyone that tens of thousands of machines all over the world can be connected to work together to maintain a ledger.

After finishing Ph.D. in 2017, Fan decided to devote himself to blockchain research, and realized that the blockchain and cryptocurrency world is no longer fun or revolutionary.

"Casino", Fan gave a new definition for the current cryptocurrency domain

"Swap CNY, US dollars to BTC, ETH or other chips-like cryptocurrencies from the exchange, and then play a variety of games like 1CO, IEO, High Multiplier Conversion, or gambling Dapps on Tron and EOS directly." Fan said bluntly, "When the player is satisfied, go to the exchange to swap the chips to fiat, or to stable coins if only wants to leave temporarily."

In fact, since ZhenFund Founder Bob Xu’s “calling for everyone” last year, many elites and academics have entered the blockchain and cryptocurrency field. However, in the general public’s perspective, Blockchain is still labelled with MLM, Scam or worse.

Hence the question, why Conflux enters such a field?

"The most intriguing part about research and technology is to find a meaningful direction for the future, and you can really contribute something." Fan said, "and it is irrelevant to achieve financial freedom or how much money one could earn from it.”

It's rare to find such an outstanding team like Conflux even globally .

“Our core developers mainly come from three circles,” Fan introduced, “Tsinghua Yao Class, academics, and frequent contest competitors. Contests like International Information Science Olympics Competition (IOI), ACM Competition, we have a dozen of gold medals within the team.

"If you intent to gather a group of people to make technological breakthroughs in the computer field, these three circles should be the most reliable source." Fan told 31QU.

the Misunderstanding of the Impossible Triangle

“Be the first public chain that achieves high TPS while not sacrificing any degree of decentralization or security.” This is a brief introduction on Conflux's official website.

In fact, current public chain technology is confronted with various problems in the course of commercialization and adoption, with the Impossible Triangle - extensibility, security and decentralization cannot be achieved simultaneously - comes first and foremost.

Therefore, emerging public chain projects that claim to be able to break the Impossible Triangle are exceptionally popular, and the on-chain and off-chain scaling solutions for Bitcoin and Ethereum chain are also discussed fiercely.

Interestingly, although Conflux claims that “the scalable consensus algorithm, with the self-developed Tree Graph, makes the consensus no longer the bottleneck of Blockchain performance”, solving the Impossible Triangle problem has never been an emphasis.

"Some Blockchain practitioners still hold bias towards the Impossible Triangle," Fan told 31QU, "The Impossible Triangle problem definitely exists, however, it is rather a problem of common sense than of technology."

Here allow us try to explain the Impossible Triangle in layman terms. If a Blockchain targets to be perfectly safe, then it requires as many practitioners as possible to participate in the transactions. However, due to the hardware differences among the participating machines, and the limitations of each machine’s processing power, along with the increasing of con-current transactions, the overall speed of the network would decrease. Therefore, the Impossible Triangle problem.

The common solution of the Impossible Triangle is to either limit the number of participants or to limit the amount of information the network processes. One example is sharding, which could break the limitation of "everyone-verifies-all-transactions" but at the expense of partial security.

"But one point everyone ignores is how much a machine can handle." In Fan’s opinion, many people deliberately avoided this question but have a big talk about the Impossible Triangle instead.

To answer the question of "the limit of one machine's processing power", we need to answer another question first: Bitcoin can only process seven transactions per second, and Ethereum can handle more than ten, what exactly is it that limits their speed?

The answer is consensus.

“If all participants intend to process the transactions together, firstly, everyone has to agree on the order of transactions.” Fan explained, “This process is achieving consensus.”

The processing speed of Bitcoin and Ethereum is limited because of the consensus protocol limitation - the algorithm to determine the transaction order is fairly slow.

Actually, this is the first breakthrough of Conflux: breaking the bottleneck at the Consensus level.

After determining the order of transactions, all the participants in the network follow this order. However, the execution process is truly the test of the processing power of each participating machine.

When dealing with each transaction in the network, the CPU processing power, the network transmission and forwarding power as well as the hard disk reading and writing power are all involved. At present, the capacity of an ordinary computer is far beyond 7TPS or 15-20TPS of the Bitcoin or Ethereum respectively .

"In our internal experiment, we re-ran all Ethereum’s first year transactions, in total several millions of transactions, on the Conflux chain. Conflux finished all of them within 1 hour with 2000-4000 TPS."

"If the target is 100,000 or 200,000 TPS, the Impossible Triangle challenge will be encountered. But for a few thousands of TPS, the Impossible Triangle challenge doesn’t exist." Fan explained, "Given the system structure is designed correctly, and the protocol is optimized enough, what is the limit of a common computer?"

"Before hitting that limit, I don't think it’s necessary to discuss the Impossible Triangle." Fan told 31QU.

The dispute between PoS and PoW

Different from most of the peer public-chain projects, Conflux insists on PoW, which is considered an expensive "consensus" from others’ perspectives.

But Fan corrected this point: PoW and PoS are not consensus algorithms; they are anti-Sybil Attack mechanisms.

"Nakamoto consensus and Byzantine consensus, they are 'consensus algorithms'. PoW and PoS are voting mechanisms within that consensus algorithm." Fan explained.

For example, the consensus algorithm decides the form of the meeting, and PoW or PoS decides whom to participate in the voting.

The first point that needs to be clarified is that there is no “identity” on the Internet. For instance the huge amount of shadow accounts in the online forums, and it is impossible to determine who exactly it is behind those accounts.

The lack of identity has caused a problem: on the internet, how to define "one person, and how do we define "majority agrees"?

In order to realize anti-Sybil Attack, which is to prevent a malicious node has multiple "identities", the designer needs to find something that cannot be largely counterfeited. In this case, PoW finds the computing resources, PoS finds the money and Stake in the system.

Secondly, Fan believes that although many people advocate "Blockchain is omnipotent" and "Blockchain changes the world", but at present, the real Blockchain can't do anything. "What the Blockchain can do right now, either privacy or transaction, can both be done better by a centralized system."

“The only value of the Blockchain is that the end results from a blockchain system are guaranteed by the credibility of half of the participants.” Fan explained, “The Blockchain will be executed in accordance with the rules and the result is irreversible, unless more than half of the participants are playing evil. "

On the other hand, any centralized organization, although efficient, always comes with two risks, one is the center doing evil, the other is the center going down.

The Blockchain provides the value of credibility, that credibility is guaranteed by people who participate in the consensus.

This is the direct reason of why Conflux chooses PoW. "As a new public chain, I hope that the generated and validated bits from Conflux is trustworthy from the beginning." Fan said, "Therefore, I hope to have as many early participants in the consensus as possible.

“I am also very puzzled by many PoS public chains, for why the investors and the project teams are the only ones who own the stake, and all the consensus is in their hands,” Fan argues, “All their credit value added together, it may still not as trustworthy as the centralized Tencent and Alibaba. Then why don't we use WeChat pay and Alipay directly?"

However, Fan also said that he’s not supporting only PoW or only PoS: "I think both PoW and PoS have their own merits, but at an early stage of a public chain, it must be PoW."

In the meantime, the PoW protocol is currently confronted with centralized mining pool, computing power competition and resources waste etc. problems.

"Many problems are caused by the mechanism design of the PoW protocol." Fan told 31QU. For example, Bitcoin only generates one block every ten minutes, and the probability of generating the block is like winning the lottery. In order to secure the profit, the mine pool has formed.

Another breakthrough of Conflux is creating a high-speed PoW protocol.

"We are fast in block generation. On the latest testnet, we generate one 4M block per five seconds. We hope to achieve four 200K-300K blocks per second in the next version.”

Besides avoiding the “winning the lottery” situation, certain game theory design in Conflux's mechanism also prevents unproportionate advantages of large mining pools against small mining pools and against individual miners.

"For example, currently mining pools charge 1% - 2% handling fee. If the advantage of joining the large mining pool is less than 1% - 2%, the miners will not be incentivized." Fan told 31QU.

From DAG to Tree Graph

If you have been following the Conflux project, you'll notice that Conflux moved its technical focus from "DAG (Directed Acyclic graph)" to "Tree Graph" in recent publicity.

"We are not just a simple DAG. We have a tree that integrated with the DAG." Fan explained.

Each Conflux chain block contains two kinds of edges, a parent edge and a reference edge. The parent edge and the reference edge together form a directed acyclic graph. But if only look at the parent edges, they forms a tree.


"We haven't changed in technology, just switched to another name" Ming Wu, Conflux CTO explained, “Many people compare us with other DAG projects, so we highlight the difference from the name.”

"A lot disputes are going on, claiming that your DAG structure is the same as mine, is it plagiarism?" Fan seemed helpless when talked about is, "What can be copied? In the absence of an attacker, in whatever way to draw the structure graph, it is right."

Fan is often asked about the difference between other DAG-based projects and Conflux, every time he would answer bluntly, "The other one is wrong, because there will be problems as soon as the attackers come."

On the other hand, the Wooden Barrel Theory also applies to the public chain. To improve the overall performance to the level of thousands, a careful design on each layer including the consensus algorithm, the network layer and computing layer, needs to be carried out. “This is the hardest part.”

"Conflux cannot raise to a hundred times in a month"

It is common to justify the overall quality of a project by its coin price in the blockchain space. It seems like nothing can be solved by an up-rising K-line. As long as the coin price rises sharply, all the flawed or exaggerated technology can be ignored and forgiven by some investors.

"It’s good." Fan thought, "Everyone demands differently. If the expectation is to buy a coin that is able to appreciate a hundred times next month, you should work hard to explore and find High Multiplier Conversion coins or similar , and strive to get rid of it before collapsing.

"We Conflux probably won't appreciate a hundred times in the first month. I’m fully aware of that." Fan smiled.

Fan believes that currently the window for public chain startup has closed, "because it is difficult to drive another public chain by capital now."

The year of 2018 was once known as the First year of the public chain. After the 1CO chaos, assorted public chains with shiny backgrounds debuted, and officially launched their mainets in recent quarters.

Conflux also comes with spotlight, but Fan disagrees that the field of the public chain is competitive.

“The competition in the blockchain industry is very small.” Fan told 31QU, “The blockchain space is a complete blue ocean. Everyone has no map and everyone’s heading towards different directions.” Only no one knows if their direction will take them to the treasures in the field.

In Fan’s words, "Conflux" is the name out of a whim.

Conflux originally means “a lot of things come together”. Fan likes this explanation, "A blockchain platform should bring all the valuable information together."