The Art of Ethereum Design (Part 1)

As the first phase of Ethereum 2.0 approaches, those who once thought that ETH2 would never come have begun to pay attention to the development of Ethereum again. On October 19, 2020, Bankless invited Vitalik to have an in-depth discussion on the design concept and original intention of Ethereum.

At the beginning of the interview, Vitalik briefly reiterated the ETH2 roadmap and said that as advertised, in Phase 0, a basic PoS public chain that anyone can use will be built, and in the first phase, sharding and Rollups will be provided for application expansion, eventually reaching 100,000 TPS.

Why PoS+Sharding

The application of a complex system like PoS is mainly to achieve several key goals. One is to eliminate the inefficiency and energy waste of the PoW system, and the other is that as ASIC becomes more and more mature, the PoW system needs to avoid centralization risks. Compared with the PoW system, the PoS system is more democratic and open, and people can participate in the network more easily, whether they are ordinary users or pledgers.

Sharding is a scaling solution whose main purpose is to increase the total transaction volume that Ethereum can handle. Scalability is critical for Ethereum because Ethereum should be a global public infrastructure that anyone can interact with. As the underlying infrastructure of the new era Internet, it should be open and easy for anyone to participate. If decentralized scaling solutions such as sharding are not adopted, the alternative may be to allow centralized institutions to run a super network and then connect to trusted side chains, but this is not what Vitalik wants to build. There have been many historical events that have proved that over-centralized systems are very easy to manipulate and will eventually deviate from the original design intentions of the designers.

Keeping the original intention of Ethereum

Decentralized permissionless network

The Ethereum network does not want to be a network that relies on super nodes, which is what distinguishes the Ethereum network from other networks. Without relying on any supercomputers, Ethereum should be a fully functioning system. The ideal situation is that a group of home laptops are running the Ethereum network. As some Bitcoin believers have said, if the network is overly centralized and controlled by a few institutions, it may eventually go in an unsatisfactory direction.

EHT2 wants to create a PoS ecosystem where anyone can participate, which is why Ethereum has never been inclined to the super node approach, nor does it believe in the assumption of "majority honesty", which is very similar to the idea of ​​Bitcoin. In the network, participants are better off verifying the chain themselves and try not to trust miners, because miners are not necessarily honest.

Overall:

1.ETH2 is more environmentally friendly and more efficient;

2.ETH2 ensures that most ordinary users can directly participate in chain writing;

3. At the same time, it ensures that most ordinary users can directly read the chain data without relying on trusted APIs;

4. Allow most ordinary users to participate in the chain consensus.

If 2, 3, and 4 are combined together, we get a blockchain that is highly resistant to audits and sociopolitical attacks.

Do people really care about these qualities that Ethereum upholds?

There are many Ethereum competitors that compromise on decentralization, permissionlessness, and Ethereum's conservative nature, but the centralized path is essentially very efficient in the short term, but it will have serious side effects in the long term. For example, a typical example is Steemit, which was acquired by Justin Sun. Excessive centralization will eventually lead to community forks. EOS has also experienced bribery attacks. Ultimately, the core participants in the centralized chain will collude and erode the rights of ordinary users.

From a governance perspective, if the voting and governance rights of a chain are controlled by large coin holders, this will be very dangerous in the long run.

Therefore, in the long run, users are very concerned about decentralization, permissionlessness and other characteristics. It takes time for people to gradually realize the importance of these characteristics. When the overall environment is good, there is no dispute in the community, and everyone's opinions tend to be unified. In this case, everyone is happy and there will be no problems with governance. But when problems arise, disputes arise in the community, opinions are no longer unified, and the rights and interests of users of overly centralized chains may be damaged.

Other communities that uphold these qualities

Coda is good. They are trying to build a chain that can use zero-knowledge proofs. In this way, zk snarks can verify the chain directly on behalf of individuals, and nodes cannot do evil directly. Mimblewimble and ETC are also good. Although ETC does not implement sharding, it tries to maintain a decentralized expansion plan.

However, for projects that stick to these characteristics as a selling point, they are not as popular as projects that focus on other concepts. Because sticking to these characteristics is bound to mean that the stories that can be told are very limited, while if you tell stories with other concepts, you can tell a lot of things. Although there will be few chains that focus on these characteristics, Vitalik believes that these projects will eventually be very successful.

PoS and Sharding

Before the release of ETH1.0, PoS and sharding had already been identified as the direction for improving ETH. Vitalik answered the question of how the community should choose the technical direction in the early stage.

As early as 2015, the community had basically reached a consensus on the technical path of PoS and sharding. The occurrence of the DAO fork further consolidated this. Those who opposed PoS and sharding at the beginning also opposed the DAO fork, so they finally chose to support ETC.

It took a while to be sure that PoS and sharding were the right technical direction. Looking back at the first blog about Slasher, you will find that at the beginning of this article it is written: Slasher is preparing for the possible implementation of PoS in the future. In January 2014, Vitalik was actually not sure whether POS had any fundamental flaws.

If we look back at sharding, Vitalik believed that the application of sharding technology in blockchain was also an unresolved problem. In this blog, sharding was listed along with several other issues as problems that blockchain could not solve at the time, or that could be solved but required fundamental trade-offs.

After 2014, another issue became a hot topic, which is the "Nothing at Stake" problem of the PoS system. Vitalik answered it in this article. Although PoS cannot have exactly the same characteristics as PoW, PoS can absorb the advantages of the PoW system as much as possible.

Between 2016 and 2017, people conducted in-depth research on other consensus algorithms and finally figured out how to apply the security model of PoW to PoS.

For sharding, the main problem is how to choose the best security model. In 2015, Vitalik and the technical community began to try random sampling, which was a breakthrough in applying sharding technology to the chain. In 2017, the community completed a breakthrough in data availability proof, which allows the shard chain to remain secure even in the case of "majority dishonesty". Through these explorations, we found that sharding is a feasible solution.

When ETH1.0 was released in 2015, we thought that PoS would be realized in one to one and a half years, but real-time results showed that the ideas at that time were too optimistic, but everyone was very confident at that time.

Token distribution mechanism from the perspective of Ethereum’s monetary policy

The community's consensus on Ethereum's monetary policy was different from what it is now. If you look at the original Ethereum white paper, you will find that the initial issuance plan was to issue 16 million ETH per year, and to increase the issuance forever. The basic principle of this policy is that the system needs to continue to increase the issuance to maintain the security of the system. At that time, the community believed that the security cost of PoS was not necessarily lower than that of PoW, so it was possible to maintain the PoW mechanism forever. And the community hoped that not only those who participated in the network now could get ETH, but also those who participated in the network later should be able to get ETH. This was the design logic at the time.

Later, Vitalik and the technical team believed that although PoS was necessary and feasible, PoW was likely to be the most equal and fair distribution method. From 2010 to 2013, PoW was considered the most democratic token distribution model, and even became a major selling point of BTC at the time: as long as you turn on your computer, you can get some BTC. This may also be the most attractive point of BTC at the time. But in fact, PoW cannot form a long-term stable equilibrium in technology. When mining is very profitable, special mining machines will appear. BTC is the best example, from the GPU revolution to the FPGA revolution to the ASIC revolution. For Ethereum, although its PoW algorithm has a special resistance to ASIC, thereby ensuring the fairness of ETH mining, as time goes on, the GPU-based system will become more and more inclined to money and resources, and may eventually cause ETH HASH to suffer ASIC risks.

So PoW will move towards ASICs anyway. Although ETH uses GPU mining, professional ETH miners will still drive amateur miners out of the market.

This is a problem with the PoW mining mechanism. BTC already has the problem of plutocracy, and it is only a matter of time for ETH. No matter how the algorithm is improved, it cannot avoid moving towards ASIC.

The challenge in designing a token distribution mechanism is to ensure its neutrality. The advantage of the PoW system is that people know what the algorithm is like, and anyone can verify it and participate. But the distribution method like Ripple through social media is very unfair. First, not everyone can participate in it. Secondly, hackers can register 10,000 accounts to get tokens. You don’t want to see a token distribution model of the underlying public chain like this. For the PoS system, its essence is to redistribute tokens to token holders and let token holders maintain network security. In addition to the existing PoS and PoW token distribution mechanisms, there is no more neutral token distribution mechanism. If the PoW and PoS distribution mechanisms are slightly changed, there may be other disputes.

PoW vs PoS

Many people believe that the disadvantage of PoS is that as long as you participate in the network at the genesis and get tokens, you can participate in the network as a pledger forever, but miners in PoW need to keep updating their hardware facilities.

Vitalik believes that POS is not 100% better than POW. From this point of view, PoW is better than PoS in reducing long-term capture.

First of all, the PoW market is still very young and may be often affected by the outside world, but these problems may no longer be a problem for the mature PoW market. But regarding the thermodynamic limit problem often discussed by BTC miners, when the cost of producing a hash answer is as low as the limit, the efficiency can no longer be improved, which will indeed become a problem that PoW cannot solve. Secondly, for POS, if you have 32 ETH, you can become a validator yourself and get more tokens. If you don’t have so much ETH, you can also get ETH by staking with others. In the PoW system, you must have enough start-up funds to buy a mining machine, otherwise you will not be able to mine. In this way, PoS also eliminates the problem of excessive concentration of wealth. Finally, the return of POS is generally lower than that of POW. The more people participate, the lower the return.

So overall, PoW has its advantages, while PoS is easier to participate but the returns are lower.

The most important point is what to do if someone owns 51% of the coins or 51% of the computing power. Compared to PoW, PoS gives the community more rescue plans. For POW, you can only watch the 51% attack happen. Maybe you can soft fork, but the attacker can continue to attack the forked chain. Vitalik calls this attack "SPAWN CAMP ATTAC". To solve the problem, you can only change the POW algorithm, but this not only hurts the attacker, but also hurts ordinary miners. When the PoW algorithm is changed, no one will have ASIC mining machines with the new algorithm for a period of time. If the attacker monopolizes the CPU GPU market, the attacker can continue to launch attacks, and then the chain will be over. So for the PoW chain, he can't do anything about the 51% attack. But it's different for the POS system. Those who are attacked can restart another forked chain and delete the coins obtained by the attacker on the forked chain, and you don't need to hard fork. So the attacker will lose a lot of coins every time he attacks.

Why should you participate in ETH2 staking?

Vitalik sees this as an announcement as an Ethereum citizen, and also as an incentive to receive income.

If you cannot be a validating block node, you should also try to participate in the network to verify the Ethereum network. The reason is mentioned in this article. You don’t have to verify everything on the chain. You can selectively verify, such as verifying data availability, or verifying fraud proofs, or even running a light node to avoid trusting a certain node. This autonomous behavior is not only good for the users themselves, but will also eventually benefit the entire Ethereum ecosystem.