Following a $4 million seed round earlier this year, blockchain protocol Algorand announced another $62 million in equity funding this month.
Regardless of the ICO that was full of bubbles before, this is still a lot of financing in the field of public chains. The amount of financing is not only huge, but 32 investment institutions including USV, China Merchants Bank International, and Xinyi Capital (See the official website for details). What's more, the blockchain is in a long winter, the project is broken, and the exchange is unsustainable. In fact, two months ago, the circle had already heard about this financing, which was also the hottest time for Algorand.
Why can Algorand raise 420 million yuan?
Angel investment is nothing more than looking at the team. Founder Silvio Micali is a professor at MIT,Co-proposer of zero-knowledge proofs, won the Turing Award in 2013 for his outstanding contributions in the field of cryptography. Due to the outstanding academic achievements of the founders, Algorand has attracted countless attention since its birth. A large number of articles analyze and analyze its consensus algorithm, and the logos of investment institutions can fill a screen. Even though there have always been many doubts about the entrepreneurship of academic teams, institutions are still unwilling to miss this project.
"Impossible Triangle" is the most talked-about technical bottleneck of the public chain in the industry, and it is also a problem that Algorand is committed to solving. Just solving the "impossible triangle" cannot make it a next-generation public chain. In addition to technology, the design of incentive mechanisms, community and talent building, and appropriate application cases are issues that are placed in front of every public chain team.
Recently, Odaily interviewed Jing Chen, Algorand's chief scientist, about these issues, and talked about Algorand's latest progress and future plans. Dr. Chen Jing graduated from MIT, mainly researching distributed ledgers and game theory, and is currently an assistant professor at Stony Brook University.
secondary title
The key to speed: local secret lottery, fork resistance
First of all, let's talk about the latest technological progress. Given that Algorand is almost a well-known project in the field of public chains, familiar readers are recommended to skip this part.
For a public chain, no matter whether it is positioned as a digital currency bookkeeping or an application platform, it needs to solve the underlying consensus algorithm, which is also one of the limitations of the scalability of the public chain. This is also the original intention of Algorand - to solve the problems of the Bitcoin consensus mechanism, that is, PoW, including congestion, energy consumption, and the centralization of computing power.
Under Chen Jing's explanation, Algorand's block generation process is not complicated. Simply put, it is to select a small group of people to run a high-speed Byzantine agreement to reach a consensus on the block. Normally, each round goes through a three-step draw:
In the first step, dozens of block proposers will be selected by local secret lottery on the whole network, and they will package the block and publish it together with their lottery results.
In the second step, the verifiers selected by lottery on the whole network will rank the nominees of the block, and find the person with the smallest lottery result as the block producer of this round. It is enough for the majority of verifiers to reach a consensus, and a Byzantine agreement will run among the verifiers. Chen Jing said that the expected number of voting users is 1000, but the specific number system will be evaluated based on the number of nodes and security. How many people are selected can make the risk probability low enough, and the same is true for other steps.
In the third step, the nodes selected by lottery will make a final confirmation of this block to determine whether there are enough users to vote for it, and accept it if it is, otherwise it will be empty.
Reduce Communication Cost and Power Consumption with VRF
This articleThis article）。
Simple random lottery is not enough for the blockchain, otherwise it is easy to encounter Sybil Attack (Sybil Attack). An attacker can control any number of public keys on the Internet. If each public key has the same probability of being selected, most of the selected public keys may be malicious. Therefore, weight considerations need to be added - the probability of being drawn each time is related to the number of tokens held by the user (Algorand's token is temporarily named Algo), as long as the malicious stake does not exceed one-third, it can run normally . It's a bit like PoS, and Micali even thinks it's the purest form of PoS.
There is actually a lottery in PoW, which is often said to solve the problem-because the solution to this problem can only be guessed randomly, so whoever guesses more times (computing power) in the same time is more likely to guess correctly, and it is not sure There are lucky people who guess a few times and hit it. Competing in computing power makes energy consumption very large. VRF's lottery program has very low power consumption and runs locally, and there is no need for communication between nodes, so the speed is very fast. Chen Jing said that even an ordinary laptop can be a node.
Resistance to forks
In order to prevent forks, Bitcoin limits the block generation time to 10 minutes, and the transaction needs to be confirmed six times. It is difficult to slow down the transaction. Algorand's protocol is resistant to forks. Even if the network is divided into two halves, the communication between the two parts is completely controlled by the attacker and will not fork.
Assuming that the system is casting the 100th block, due to the existence of malicious nodes, for example, the block proposer sends different blocks to different users; or the attacker divides the network into several parts and temporarily controls the communication between users.
In different attack scenarios, there may not be any block with sufficient number of signatures in the third step above, or there may be a block with sufficient signatures but the attacker hides these signatures through the control of the network. An average user cannot immediately distinguish the two cases. At this time, the system will enter the second cycle of generating the 100th block.
Chen Jing said: "At this time, you have to vote for the new block No. 100, and you must be very careful. Especially in the second case, you must ensure that the block generated in the second cycle is the one that has enough signatures before. "
The reason why Algorand can not fork under extremely strong network attacks, Chen Jing explained, is that the method for each user to vote in the subsequent cycle depends on what he sees in the previous cycle; A sufficient number of conflicting signatures can be avoided in parameter design.
secondary title
Difficulties beyond the "trilemma": sustainable development
Impossible triangle, also expressed as trilemma (trilemma). "Although it is called a trilemma, there is no strict proof of impossibility." Chen Jing does not think that this triangle cannot have both. Aside from the controversy over the existence of trilemma, we talked more about designs other than technology, such as incentive mechanisms, community building, and talent recruitment.
Sustainability, this is the word that Chen Jing mentioned the most in the interview.
From fundraising, announcement of financing, design of incentive mechanism, to team recruitment, the team hopes to do this.
The reason why I did not choose ICO but chose equity financing is to consider sustainability; the same is true for choosing the time to disclose financing and team information, and plans to disclose incentive mechanisms and other technical documents in the future.
Dangerous Incentives: Bitcoin's Alienation
"Motivation is the hardest thing." Micali in 2017Malta's Financial Encryptor Data Security Conferencesaid at the first presentation of the Algorand mechanism.
The lack of an incentive mechanism is the biggest controversy. As Coindesk reported, Micali said at the time: “We have to use incentives as a last resort. I believe I can [make algorand work without incentives], but I don’t have the exact proof that I can.” Many people think that Micali is too idealistic, and Algorand has not introduced an incentive mechanism or issued digital encrypted currency so far.
In Chen Jing's view, rather than saying that Algorand did not design an incentive mechanism because it was too idealistic, it is better to say that they have not released an incentive mechanism because they are too cautious. Since Algorand's consensus protocol is not based on PoW, the cost for users to participate in the consensus protocol is very small. This makes the incentive mechanism less decisive for the consensus protocol itself than it is for the PoW system. But the ecosystem based on the Algorand blockchain needs an incentive mechanism, which the team has made clear from the very beginning.
"Actually, we have an incentive mechanism at the beginning of the design." At the same time, Chen Jing also admitted that it is too difficult to design a suitable incentive. "Just like Bitcoin, once there is an incentive mechanism, you don't know how users will try their best to exploit various loopholes to make money."
Bitcoin, the public chain with the largest market value in the world, can be described as an incentive for success and incentive for failure.
At the beginning, the computing power threshold of Bitcoin was also very low, but the incentive mechanism turned democracy into oligarchy. "Bitcoin's current situation is contrary to decentralization." She believes that a large amount of computing power is concentrated in mining pools, and the two groups of miners and ordinary users are very inconsistent. Micali also expressed similar sentiments.
Code cannot solve everything, incentives need to be both online and offline
Algorand does not want to centralize the share of currency holdings, otherwise currency holdings will become another kind of monopoly.
The lesson of Bitcoin is that code cannot resolve human nature. After studying the distribution of tokens for a long time, Algorand, a team at the forefront of technology, not only relies on codes, but also resorts to offline means to "prevent the centralization of votes (stake)".
1. Hope to help community building through incentives, including developer communities and user communities. "The members of our team are very affectionate to the school they used to be. We value the cooperation with the university, whether it is as a user or the distribution of tokens, it is very important." Algorand will invite top universities around the world Become a node, do research and application in the network, and the Algorand Foundation will allocate some tokens to the university program. “We think that universities attach more importance to social welfare. They do research on it, which may inspire everyone to do more and more in-depth exploration of the blockchain. At the same time, universities are a good way to connect with the community.”
2. Online participation is not necessarily rewarded, and offline behavior may also be rewarded. In Bitcoin, there are only rewards for bookkeeping, which leads users to continuously increase their computing power in order to compete for the right to generate blocks. In order to avoid repeating the same mistakes, Algorand will reward the running program itself, regardless of whether it is drawn or not, it may be rewarded, a bit like a personal version of cloud computing vendors. At the same time, participating in community building and activities, application development and other offline behaviors may also be rewarded. Part of the algos held by the Algorand Foundation is dedicated to community building, token design, etc.
3. Set up an auction mechanism to maintain community attention. "We don't want to do an ICO and sell out the tokens at once, and then everyone loses attention, and we can't provide long-term effective reference for price discovery. We will distribute some tokens through auctions on our mainnet , such as one auction per day, for 5 years.” I feel that this is a bit similar to the year-long ICO of EOS. The difference is that Chen Jing told me that the Algo auction will be synchronized with the launch of the main network and will not be in advance; and the auction will be held on the main network for a long time.
She didn't reveal too many details about future incentives. In addition to considering the release rhythm, the incentive mechanism is too easy to be copied is also one of the team's concerns.
text
Solve the consensus mechanism first, and then add smart contracts later
The current Algorand is still a "distributed ledger", that is, it can only transfer money, without adding functions such as smart contracts. The test network has been launched, and the main network is planned to be launched in 2019.
Its first version is mainly to solve the consensus problem, to ensure that there is an underlying decentralized architecture, and there will be no full-state virtual machine. After that, the platform will gradually launch new functions according to industry needs, such as digital asset functions for the financial industry. As the number of platform applications increases, the circulation value of tokens increases, and a more stable value may also be required. This is what the team will consider in the future.
Reference article:
I am Lu Xiaoming, editor of Odaily. I am exploring the real blockchain. Please add lohiuming for breaking news and communication. Please note your name, unit, position and reason.
Reference article:
MIT Professor and Turing Award Winner Silvio Micali Raises $4M to Build a Better Blockchain
Counter-intuitive Algorand: breaking through the "impossible triangle", Utopia or Utopia?
No Incentive? Algorand Blockchain Sparks Debate at Cryptography Event
Goldwasser, Micali Receive ACM Turing Award for Advances in Cryptography
Algorand Releases First Open-Source Code: Verifiable Random Function
Algorand: Scalable Byzantine Agreement for Cryptocurrencies
