The night before the launch of the main network, we talked to the core team of Algorand about the road to centralization.

This year is destined to be a key year for the block chain industry. We not only need to meet the march of giants such as Facebook, on the other hand, the third generation of public chains after Ether Square, including Algorand, Polkadot, Nervos, Conflux, etc., have also arrived at the main online line one after another. The public chain, the most important infrastructure in the industry, was almost entirely smashed into the track in 2017. Now, the time has come to test the results.

Title: "the third Generation Common chain of early delivery, the atypical decentralization of Algorand" by Orange Book Retric

Title of the original text: "the third Generation Common chain of early delivery of papers, the atypical decentralization of Algorand"Author: orange Book Retric

Algorand is currently the first project to hand in a paper. They will officially launch on the main network on the 19th of this month. Along with the main network, there is a five-year token auction plan: through a Dutch auction, Algorand will systematically distribute Alog, the original token of the POS system. The auction will take place on the compliance platform Coinlist, and users will need to participate through KYC.

Unlike POW, which digs coins through mining, the security of the pure POS system depends largely on the initial issuance of tokens. If the initial release of the token is dispersed enough and the money holders are decentralized enough, the system will be more secure.

To this end, Professor Micali, founder of Algorand and winner of the Turing Prize, believes that Dutch auctions are a more suitable way to distribute token than ICO and IEO. At the same time, behind this way of issuance, it is also the first attempt by Micali to realize the democratization of finance with Alogrand.

It is also clearly a bold and creative attempt: a five-year continuous auction, two times a month, with a monthly target of 50 million Algos,90% guaranteed refundable mechanism. Among them, the 90% guaranteed refundable mechanism is undoubtedly a design bright spot, the users participating in the auction have a refund opportunity a year after the auction, depending on the original auction price, up to 90% of the original amount can be returned.

This particular auction makes the token issuance process full of games, and while the refund mechanism ensures the security of the investment, it may also create a lot of potential secondary market arbitrage: for example, after the auction, buy when the token price falls, and then choose to refund the original auction price a year later.

The Dutch auction mechanism itself is unusual. In general, the auction price is from low to high, but the Dutch style is just the opposite, the price will be from high to low. In the process of falling prices, you can participate in the auction, when the price drops to the total amount of the current auction pool is just enough to buy all the tokens of the day, the auction is over.

For example: suppose a total of 100 ALGO tokens are sold one day. Starting at $10, Xiao Li offered $50 for 10 at $5. After that, the price continued to fall, and Xiao Wang offered $40 for 20 at $2. After that, the price continued to fall, and when the price dropped to $1, Xiao Ming offered another $10 for 10. By this time, there was already $100 in the auction pool, and the auction price had dropped to $1, which was exactly the price of 100 ALGO. At the end of the day, Xiao Li won 50 ALGO, Xiao Wang won 40 ALGO, and Xiao Ming won 10 ALGO. Although Xiao Li made the offer at $5, his final transaction price was actually $1.

There are not many POS projects online in themselves, Cosmos is one, but the token issuance model of Cosmos is very simple, just using the original private placement. In view of the fact that token issuance is very important to the POS project, and many people in the industry have put forward the latest issuance mechanism, including Lockdrop, in this context, Algorand, as a star project of pure POS, has creatively used the Dutch auction + 90% return capital preservation mechanism for as long as five years to spread the coin, it is not difficult to imagine that if this way has achieved good results, It will have an important impact on many later POS projects. At least POS star projects will consider following this release.

With the exception of the token issuance mechanism, Algorand is independent in many ways.

With the exception of the token issuance mechanism, Algorand is independent in many ways.

You can clearly feel the difference between it and other public chain projects. Although the starting point of each project and the stories they tell are very different, the core motif of the third generation public chain is the same, that is, on the basis of ensuring decentralization and security, expand the performance of the blockchain as much as possible. The same is true of ETH2.0 from POW to POS.

In contrast, Algorand has adopted a different design in the face of this proposition: it has abandoned the creative incentive mechanism introduced by Satoshi Nakamoto for Bitcoin as much as possible, and has returned to the cryptography system that Micali itself is familiar with, trying to use the VRF + self-election mechanism to break free of the impossible triangle.

It has abandoned the creative incentive mechanism introduced by Satoshi Nakamoto for Bitcoin as much as possible, and has returned to the cryptography system that Micali itself is familiar with, trying to use the VRF + self-election mechanism to break away from the impossible triangle.

Last October, Orange Book interviewed Micali, about this design idea, and we had the opportunity to sit down again with Professor Micali and Dr. Chen Jing, chief scientist of Algorand, to talk about some of our concerns before Alogrand went online.

What kind of mission does Algorand want to accomplish? Will the first version of the main network have a smart contract? How will the outstanding incentive issues be resolved? Why are Dutch auctions used for the initial issue of tokens? As a rookie, should I take part in the Algo auction?

After reading the following interview, some of these questions of concern may be able to get some answers, but some questions still need to be judged by you (such as whether you should participate in the auction).

The democratization of The Democratization of Finance Finance

Orange Book: the thinking of each project in the third generation public chain is obviously very different, each is solving different problems, and each thinks that the problem they are solving is the most important. What is your vision for Algorand? What kind of questions do you think are the most important to the current public chain? What kind of solution does Algorand give to this problem? What do you think Algorand can help people do that they couldn't do before?

Orange Book: the thinking of each project in the third generation public chain is obviously very different, each is solving different problems, and each thinks that the problem they are solving is the most important. What is your vision for Algorand? What kind of questions do you think are the most important to the current public chain? What kind of solution does Algorand give to this problem? What do you think Algorand can help people do that they couldn't do before?

Micali: what I want most is that Democratized Finance, uses Algorand to democratize finance. Take our Dutch auction, for example, this is the first application on Algorand, and this is one of the applications I want the most. If, in the traditional auction, you need all the participants to gather in the same room, but we put the auction process on the chain, and the chain is really extensible, then people all over the world can participate.

Take our Dutch auction, for example, this is the first application on Algorand, and this is one of the applications I want the most. If, in the traditional auction, you need all the participants to gather in the same room, but we put the auction process on the chain, and the chain is really extensible, then people all over the world can participate.

For me personally, this is my initial motivation and enthusiasm to be a Algorand. Of course, there are a lot of people who say they want to play games on Algorand and do DAPP, but what I personally want most is the initial auction app, which represents the democratization of finance.

The financial system of our present world is not so friendly to ordinary people. It is not equal and not everyone has free access. Even many very old financial products and services are not democratic. We want more complex, finer financial applications that are equal and freely accessible to ordinary people, because only in this way can you go to Elevate Everybody (to help improve people) and let everyone share the possibilities of finance.

The technical advantage of Algorand is that it removes a lot of mining costs, because Algorand is truly extensible and decentralized, so we can afford to treat everyone equally so that everyone can fairly participate in the system.

Orange Book: did this auction take place on the chain? Then why work with Coinlist at this auction? you need KYC?.

Orange Book: did this auction take place on the chain? Then why work with Coinlist at this auction? you need KYC?.

Micali:, yes. This is actually because you can't achieve decentralization in the first place. Decentralization takes a process. Let's think about the initial state of the Algorand system-no one had Algo tokens at first, and if you want to find someone to buy tokens, you must have bought them in French coins. But we can't put the French currency on the blockchain, so at first you need a role like Coinlist, which accepts your legal currency and then registers you to be able to bid. In the end, everyone's bid will be on the chain, and you can see in real time how many tokens are left on the chain. When everyone has an Algo token, we don't need a character like Coinlist. For example, if you want to make a Dutch auction of a property, just auction it with a Algo token on the chain.

Orange Book: why did you choose to do Dutch auctions on the chain? Why not ICO or IEO, or something like Lockdrop?

Orange Book: why did you choose to do Dutch auctions on the chain? Why not ICO or IEO, or something like Lockdrop?

The Micali: Dutch auction gave the pricing power entirely to the market. Not like ICO, set a fixed price. How much is a coin reasonable? No one knows. The price of Algo is not determined by me, and I am not interested in pricing it. Algo is determined by everyone in the market.

This is also an example of how we want to democratize finance. Because you want to treat every participant in the system equally. Each user enters a new ecology, he will care if he is treated equally, and is treated equally from the beginning. We also want the system to be decentralized from day one.

When is the first day? That's the first time you bought tokens. After the Dutch auction, it is relatively fair for everyone to pay the same price for token,. Not all auctions meet this requirement. Dutch auctions happen to be like this, so we chose this auction method.

Another function of Coinlist is to provide us with a refund policy. The policy is, assuming you don't want Algo, in a year's time, you can ask us to refund 90% of the money (if the transaction price is greater than $1). When you want to return it, someone has to know who you are, right? So you need Coinlist to do KYC. By the way, if you buy a car, you want to return 90% in a year. Do you think it's possible?

In blockchain systems, you want not only to treat users fairly, but also to make users feel fair. Putting Dutch auctions on the chain can do that. If you don't put it on the chain, of course, you can call your users and say, what's the price of a real-time auction right now, which may be fair, but how do they feel that? I don't think there's anything I can do about it.

Orange Book: I'm curious. A lot of the projects I've seen, including Bitcoin and Ethernet Square, are actually initiated by grass roots. Nakamoto, we don't know who he or she is, but most of the people around Bitcoin are ordinary people, and so is Etherland. Vitalik is just a very smart kid. If the blockchain is a bottom-up revolution and you (in my opinion) are at the top of the elite, why would you want to do it?

Orange Book: I'm curious. A lot of the projects I've seen, including Bitcoin and Ethernet Square, are actually initiated by grass roots. Nakamoto, we don't know who he or she is, but most of the people around Bitcoin are ordinary people, and so is Etherland. Vitalik is just a very smart kid. If the blockchain is a bottom-up revolution and you (in my opinion) are at the top of the elite, why would you want to do it?

Micali: I don't think I'm at the top of the pyramid, of course, we're better educated, you can say we're elite, but I think that's our responsibility to go to Elevate Everybody. Responsibility often comes with your abilities.

The real key problem is that we have been lacking the technology to help everyone. If, under the old system, you want to do one thing for people, you have to communicate with a small number of people, and then let that small group of people communicate with the rest of the others, then this small group of people will certainly have an advantage over others. They will have more rights. This is not equal to all people.

If you look at the Internet, it is a very democratic tool. Today, information flows at a very cheap rate on the Internet, which is the result of the democratization of information. Information can flow freely, but value and money do not flow freely because the cost of the transaction and the friction of the transaction are still high. For the rich, it is acceptable to pay part of the cost of exchanging value to avoid the friction of some transactions, but for the poor, this is impossible. When the value I want to exchange is very small, the cost of the transaction is greater than the value, then the value transfer may have been exhausted halfway through, and I will not be able to exchange value.

Today, information flows at a very cheap rate on the Internet, which is the result of the democratization of information. Information can flow freely, but value and money do not flow freely because the cost of the transaction and the friction of the transaction are still high. For the rich, it is acceptable to pay part of the cost of exchanging value to avoid the friction of some transactions, but for the poor, this is impossible. When the value I want to exchange is very small, the cost of the transaction is greater than the value, then the value transfer may have been exhausted halfway through, and I will not be able to exchange value.

What we need is a more efficient flow and exchange of value, not just a matter of payment, but also, as I have just said, a more sophisticated and complex issue of financial services. Most of these more sophisticated financial services are now done through intermediaries who have no intention of serving you when your prices are very low. They will tell you to go away because they will only leave time to high net worth clients.

So, for a really expandable blockchain, it can reduce the cost of the transaction, and these things are possible.

Every time bitcoin produces a block, the coin you have is actually inflated. Few of us can really understand inflation because our education is not intended for us to understand inflation. The reason for inflation is that each time a new block is added to the Bitcoin ledger, someone has to pay a cost for the miners, which is the cost of the transaction.

The way Algorand reduces this cost, first of all, is to reduce the cost of participation: you don't have to fight to solve mathematical puzzles, you don't have to take part in an arms race, so you can participate in the Money Trust of Algorand at a lower cost.

But Money Trust is only one thing, and you want to democratize other more complex financial products and instruments-such as lending, which is a very basic economic component. Imagine that if you want to open a hotel and don't have enough money at first, you need to borrow money to start the business. When the cost of borrowing is high, it means that the interest on borrowing is very high, and if we can reduce the cost of borrowing through the blockchain, we may be able to encourage more people to borrow at rates they can afford to start their own hotel business. That's what I want to do.

Orange Book: so the problem with DeFi at Ethernet Square is that the chain itself is too expensive?

Orange Book: so the problem with DeFi at Ethernet Square is that the chain itself is too expensive?

Micali:, on the one hand, it's too expensive, and at the same time, it's too slow, and it's obviously going to be worse when these two things combine. Its costs and trading frictions are there, unable to give more people equal access to complex and sophisticated financial services.

Orange Book: I still want to go back to that question: if the blockchain is a bottom-up movement, people like Vitalik obviously seem more approachable and closer to the community. Are you worried about how, as an elite, you can connect with a decentralized community? Does the elite still have an advantage in this respect?

Orange Book: I still want to go back to that question: if the blockchain is a bottom-up movement, people like Vitalik obviously seem more approachable and closer to the community.Are you worried about how, as an elite, you can connect with a decentralized community? Does the elite still have an advantage in this respect?

Micali: I think what matters is not whether a system is initiated by ordinary people, but whether the system can really serve (Work For Common People) for ordinary people.

A few months ago I met a core developer named Greg who said he was very loyal to Etherland and that your technology was interesting, but I wouldn't leave Etherland. So I asked him, Greg, do you think you should be loyal to Ethernet Square, or should you be loyal to decentralization? Then Greg joined Algorand, because they kept talking about Casper is coming, but Casper nerver coming. Greg wants to join Algorand, because he is concerned about whether the agreement can eventually serve ordinary people.

I think the best thing is that we're working on an open source agreement that doesn't need a license. Everyone is free to choose, and eventually the real work agreement will come out.

About University Node

Orange Book: why did you choose to work with universities to make them the first nodes of Algorand?

Orange Book: why did you choose to work with universities to make them the first nodes of Algorand?

There are a lot of things in Micali:Algorand that need to be voted by the community, but what if we don't have such a community in the first place? So we think it's a good idea to work with universities in the early stages, and let universities help us make a lot of voting decisions, such as economic model decisions, agreement technology upgrading decisions, and so on. Because universities are: 1, non-profit organizations, 2, there are many economic experts, 3, there are many technical experts, 4, universities are international, MIT accepts students from any country.

These four advantages determine that they are good candidates for nodes. Many universities around the world, including Peking University and Tsinghua University, are now our nodes. In the future, there will be more and more members of the Algorand community, and the number of nodes will increase.

Orange Book: is the university as the Relay node in the Algorand network? What kind of work do they have to be responsible for?

Orange Book: is the university as the Relay node in the Algorand network? What kind of work do they have to be responsible for?

Chen Jing: there are two kinds of, Relay Node (relay nodes in Algorand nodes) and Non-Relay Node (non-relay nodes). Relay nodes perform deduplication, signature checking, and other verification steps, and eventually repropagate valid messages to other nodes. Relay nodes generally do not participate in the block, and there is no incentive at the current stage. Non-Relay nodes are nodes that participate in network elections, and anyone can download the client to run.

Orange Book: why are universities willing to be Algorand nodes without incentives?

Orange Book: why are universities willing to be Algorand nodes without incentives?

Chen Jing: the appeal of the university is actually to have better cooperation in the field of research. This is a more important consensus of interests between our two sides in their cooperation. Because we are doing a lot of research on block chains, universities are also interested. In addition, we will also donate some Algo token, to universities, but because of policy regulation and other reasons, these token will not be given for several years, so it is more research cooperation.

About Smart contract

Orange Book: Algorand doesn't have a smart contract feature this time, right? How can developers do financial applications without a smart contract?

Orange Book: Algorand doesn't have a smart contract feature this time, right? How can developers do financial applications without a smart contract?

Chen Jing: it is mainly done in the way of layer2. Every transaction on layer1 will have a data domain, part of the data will not be dropped by the system parse. You can write the data into the transaction + mutisig (multiple signature) to develop applications.

Orange Book: put the data in the layer1 transaction, where does the application's business logic go?

Orange Book: put the data in the layer1 transaction, where does the application's business logic go?

Chen Jing: put the business logic on your own APP.

Orange Book: the equivalent of a mix of centralization and decentralization?

Orange Book: the equivalent of a mix of centralization and decentralization?

Chen Jing: yes. This is actually a consideration of our design. We think that if the smart contract is done entirely with layer1, its cost will grow very fast. At first we just wanted to make the most important things layer1, like atomicity.

Many layer1 smart contracts are designed to ensure atomicity. Two people perform an operation that either occurs or does not happen, and there is no other possibility, so as to ensure that the rights and interests of both parties will not be compromised. So we do the atomicity of layer1, and a lot of other things can be solved by layer2, such as who should pay how much between two people, how to sign, these things can be done with layer2.

Orange Book: what if layer2 has a dispute? How do I resolve this dispute?

Orange Book: what if layer2 has a dispute? How do I resolve this dispute?

Chen Jing: this is the problem that layer2's APP has to solve on its own. For example, layer2 can have escrow, with Exchagne, and several people, but the thing these escrow needs most is atomicity, so let's do this part on layer1.

Here's the idea: layer1 starts with the most critical features, and then layer1 grows more and more new features, which is equivalent to slowly synchronizing from layer2 to layer1. A lot of things are moved to layer1 behind layer2, and then layer1 has more features, and layer2 can release more power, do more, and iterate over it step by step.

The advantage of this approach is that we can better understand how layer1 should be built. For example, layer2 may slowly appear a lot of high-value, very popular, important applications, what is the commonality of these applications, we abstract it out and feed it back to the design of layer1.

Orange Book: equivalent to slowly according to the needs of the application, and then consider how to improve the design of layer1?

Orange Book: equivalent to slowly according to the needs of the application, and then consider how to improve the design of layer1?

Chen Jing: that's right.

On incentive Mechanism

Orange Book: Algorand had a very different view of motivation from the beginning and felt that it should be used as little as possible. Now the main online line, how is the incentive mechanism designed? For example, what is the incentive for nodes?

Orange Book: Algorand had a very different view of motivation from the beginning and felt that it should be used as little as possible. Now the main online line, how is the incentive mechanism designed? For example, what is the incentive for nodes?

Chen Jing: at present, there is no incentive for nodes. The first step of the incentive is reward token holder. We have always stressed that the first step is to spread out the token, so the first step in incentive design is for the early token holder.

Whether a user marks himself online or offline, the incentives he receives in the system are the same, distributed in proportion to his token holdings. In other words, regardless of whether or not out of the block, or even whether or not to participate in the block election, the reward is the same.

Because the number of early currency holders is relatively small, the early reward rate will be higher, and later as the number of currency holders increases, the rate will gradually decrease. The current incentives are mainly for early participants.

Next, when we increase the incentive mechanism, we will add the incentives related to the nodes. There are two kinds of nodes in Algorand. The main cost of Relay node is the cost of network communication.

If you look at the cost of any blockchain system, where will its Cost be generated?

It is nothing more than computing, storage, and network communication. Specifically, each system is different. For example, the cost of bitcoin, for example, is mainly focused on computing, so the main incentive for bitcoin is also given to the calculation. Because of the low requirement of computation, Algorand does not give special incentives.

In terms of storage, Ethernet Square is involved, and the problem of explosion in all its states to the layer1, state is also very serious. Ethernet Square in the storage incentive design is strange, it will give you a reward when you release the space, so that people will do arbitrage on it, such as when the ETH price is relatively low to buy a larger space, but also do not, wait until the ETH price rises to return the space to get the reward. As a result, the space is wasted and the load on the system is relatively high.

Storage overhead, Algorand can be solved through technology, I also mentioned a little bit about our VAULT system in the field today, this system is mainly used to help new users to join the Algorand system to reduce storage overhead.

In addition, Algorand also uses "aggregate signature" and other methods to compress the amount of storage required for node signatures, so storage will not be a big problem for the time being. Moreover, in the future, there will be Archive Node (archiving nodes in the network) to be responsible for storing all the information, and your query for historical data can also be done through these Archive nodes. Some nodes become a pure participant and do not involve information storage.

Another function of VAULT is to make the Algorand system do not need to trust these Archive nodes, because these nodes give you a proof when you store information, you just have to look at the proof. Whether this node is well-intentioned or malicious, it has no way to mislead other nodes to accept bifurcation. The only way to do evil is not to work and not to respond to requests sent to it by other nodes. It is enough to ensure that there are enough well-intentioned nodes willing to work in the Algorand system.

So storage is not a problem for us, and we don't think it's necessary to design special incentives.

In summary, both computing and storage can be solved by technical means, and network communication has become the only bottleneck of Algorand. Therefore, the incentive mechanism that will be added in the next step is the reward of the Relay node in terms of network communication overhead. Later, there may be incentives for behaviors such as Block Proposer and Voting.

About cost and Security

Orange Book: many blockchain systems may have subconsciously given us the feeling that the higher the cost of the system, the more secure it seems to be. For example, if Bitcoin consumes so much electricity, the higher the cost of breaking it, the higher the security of the network. Algorand just talked about, computing, storage and network communication these three aspects are relatively not too much cost, but its security is very high? Does that seem like a counterintuitive feeling?

Orange Book: many blockchain systems may have subconsciously given us the feeling that the higher the cost of the system, the more secure it seems to be. For example, if Bitcoin consumes so much electricity, the higher the cost of breaking it, the higher the security of the network. Algorand just talked about, computing, storage and network communication these three aspects are relatively not too much cost, but its security is very high? Does that seem like a counterintuitive feeling?

Chen Jing: Algorand is designed in a very different way. Bitcoin is a trade off, that has a strong cost and security. It ensures that a block is generated in 10 minutes, and the security is high enough. Since Algorand does not need to use trade off to ensure the security of the system, cost and security have no inevitable impact, so low cost should not be a good thing in my opinion?

Orange Book: this must be a good thing. In other words, the security of a system, like a castle, the higher the wall, the safer the castle, the wall is like some kind of consumption of the castle, if this consumption is very low, is it easy to conquer?

Orange Book: this must be a good thing. In other words, the security of a system, like a castle, the higher the wall, the safer the castle, the wall is like some kind of consumption of the castle, if this consumption is very low, is it easy to conquer?

Chen Jing: I think you're talking about entropy. The greater the entropy of the system, the less likely you are to be predicted. The so-called injection randomness is to inject greater entropy into the system to ensure the unpredictability of the system, so as to ensure the security of the system. To some extent, we use VRF + self-election to increase the entropy of the system and use the complexity of information to ensure the security of the system.

Orange Book: so is POW using energy from the outside to increase the entropy of the system?

Orange Book: so is POW using energy from the outside to increase the entropy of the system?

Chen Jing: POW doesn't actually increase the entropy of the system. It uses energy to ensure safety. As you said, consume a certain kind of resources to build the wall, the higher the wall, the safer the castle. Instead of consuming resources, you can switch to a random approach, a bit like guerrilla warfare, to make yourself difficult to predict, so as to ensure the security of the system. So injection randomness and POW, are two different ways.

Orange Book: what variables will affect the security of this system if the protocol of Alogrand has been determined? For example, does its security depend only on the number of people involved in the system, the proportion of good or bad people? What do I need to do if I'm going to break through this system?

Orange Book: what variables will affect the security of this system if the protocol of Alogrand has been determined? For example, does its security depend only on the number of people involved in the system, the proportion of good or bad people? What do I need to do if I'm going to break through this system?

Chen Jing: the only way is to control enough token. Of course, if you can break the password to forge a signature, the system will certainly be breached. Or if you have a lot of hash,Algorand systems designed to allow attackers to have a power of 10 to 40 hash, in this case, design the parameters of the system to meet the security requirements. What is the concept of the 40th power of 10? it will take you hundreds of millions of years to do so many hashes.

Orange Book: so the dispersion of coins is critical to the POS system.

Orange Book: so the dispersion of coins is critical to the POS system.

Chen Jing: yes, the basic assumption is that most of the token is in the hands of honest nodes.

Orange Book: is the requirement of this honest node higher than the 50% requirement of Bitcoin?

Orange Book: is the requirement of this honest node higher than the 50% requirement of Bitcoin?

Chen Jing: the theoretical value is 2/3 honest nodes. In the actual system design, such as the main network, we may require 80% to be honest nodes, because in this way you can reduce the size and number of election committees, so that the efficiency and speed of the whole system will be better. There will be a tradeoff.

Source link: mp.weixin.qq.com

Chain News only provides relevant project information and does not constitute any investment advice.

Does not constitute any investment advice.

Firmly put an end to all kinds of token issuance and hype, if you find that the article contains sensitive information, please click "report", we will deal with it in a timely manner.

Report