Ethereum founder Vitalik Buterin once elaborated on this thinking and suggestion in his blog post Making Ethereum alignment legible: Regarding the issue of decentralization and security, we must ensure that we minimize reliance on centralized infrastructure and minimize censorship vulnerabilities. To this end, we can test and evaluate the following methods: "Walk Away Test" and "Internal Attack Test".
Among them, "internal attack testing" refers to launching an attack on the system autonomously to observe how much damage it will cause in order to discover loopholes; while "exit testing" is a newer thinking tool used to check the degree of dependence of projects and network centralization. It can become a key test for evaluating decentralized projects, and can also be improved and upgraded into a risk rating tool.
Making Ethereum alignment legible, please refer to the original article:
https://vitalik.eth.limo/general/2024/09/28/alignment.html
What is an Exit Test?
If your team and servers disappeared tomorrow, would your application still work?
This is the core testing idea of "leave testing" - this is a thinking tool that can be used to evaluate whether a Web3 project, platform or protocol has real independent operation capabilities and sustainable development value.
The “leave test” is closely related to the technical philosophy of blockchain decentralization and autonomy. The following are some of the possible directions of thinking that can be derived from this test:
Project development:
If the development team is disbanded, can the project still operate independently?
Is there an active community that can take over the project after the team leaves?
Is the project's code open source and attractive for developers to continue improving it?
Are there decentralized validating nodes protecting the network, or enough community support to maintain development?
Economic model:
Does the project have a sustainable economic model?
Does the project have application scenarios for sustainable application?
Does the appreciation of assets on the project rely substantially on speculative manipulation or centralized control?
Community governance:
Is there a way for all parties involved in the project to participate fairly in decision-making?
Can the project initiate decision-making mechanisms and resolve issues without identifying key managers?
Does the project have to rely on a small number of core members for governance, or does it have a more widely distributed foundation for collective decision-making?
Why is Exit Testing Important?
If a project is too dependent on the founding team or certain key personnel to operate; if a network must rely on a fixed server to process data, then it is still centralized in nature, and the long-term viability, value, and even censorship and risk resistance of the project or network may be questioned.
The importance of "leaving testing" lies in the fact that through this thinking tool, it is possible to uncover the actual situation of a project or network's dependence on centralized infrastructure, allowing the project or network to be effectively improved. The technical philosophy it relies on is firm "decentralization."
In 2017, Ethereum founder Vitalik Buterin wrote in an early blog post on the concept of decentralization:
"Decentralization" is one of the most common words in the field of cryptoeconomics, and is often used as a direct basis to measure whether a network is a blockchain network. However, the actual meaning of this word often causes a lot of confusion and confusion.
Vitalik Buterin pointed out that when people discuss a decentralization issue, they are actually discussing three independent aspects:
Is the architecture centralized or decentralized?
For example, how many computers does this system consist of? How many computers can this system tolerate crashing at any one time and still continue to run?
Is it politically centralized or decentralized?
For example, how many individuals and organizations can ultimately control the computers that make up this system?
Is it logically centralized or decentralized?
For example, are the system's interfaces and database structures a single entity or an unstructured group? If you split the user and provider of the system into two, can they still operate as completely independent units?
As for the role and significance of “decentralization”, Vitalik Buterin also gave a clear explanation in his 2018 blog post:
Fault tolerance: Decentralized systems are less likely to experience unexpected failures because they rely on many independent components. In theory, the probability of independent components failing at the same time is relatively low.
Resistant to attacks : Decentralized systems make it more expensive to be attacked and manipulated because they lack sensitive central points. The cost and difficulty of attacking a system with a clear central point is significantly lower than that of a decentralized system.
Prevention of collusion: If participants in a decentralized system want to conspire to benefit themselves at the expense of other participants, they will pay a higher price than participants in a centralized system.
Core Value: A Key Test for Evaluating Decentralized Projects
If we look at it from the logic of the "exit test", Bitcoin can be considered to have passed this test: the public does not know where Satoshi Nakamoto is, but Bitcoin can continue to develop by relying on a decentralized network and global developers.
In Ethereum, founder Vitalik Buterin mentioned in a forum in 2022: Currently, almost all Rollups are not mature, and most of them use auxiliary means called Training Wheels to ensure operation. However, the auxiliary means of Training Wheels also reflect the Rollup project's reliance on "human intervention". The lower the risk of the Layer2 network that does not rely on Training Wheels, the higher the risk of the Layer2 network that relies more on Training Wheels.
To this end, Vitalik Buterin and others classified the degree of Rollup project dependence on Training Wheels: Stage 0 (complete dependence), Stage 1 (partial dependence), Stage 2 (complete abandonment). Subsequently, the L2beat website revised this classification scheme by soliciting opinions from the community, and upgraded it to the "Layer2 Risk Rating Index" in June 2024 to rate the risks of different Layer2 projects.
“
learn more
What are Training Wheels?
Training Wheels (common translation: training wheels) are certain restrictive mechanisms or measures added in the early stages of Rollup technology implementation to ensure safety and stability.
The Rollup protocol that needs to implement Training Wheels has usually not yet achieved de-trust or trust minimization. The main reasons may include that the code is too complex or has not yet been audited for security, and the potential attack surface of the contract is large; the protocol has just been launched and user trust has not yet been established.
In this regard, Vitalik Buterin pointed out that his ideal goal is to see more entities like L2beat emerge to track the actual situation of various projects in meeting established standards or other standards proposed by the community. The competition between projects will no longer be "whether you have the right friends", but to "stay consistent" as much as possible according to clear and easy-to-understand standards.
From a broader perspective, "Leave Test" can actually be improved and upgraded into a risk rating tool to measure the decentralized nature and development sustainability of various decentralized use cases such as Web3 wallets, games, DeFi, etc.
As a common political philosophy theory says: the best way to solve the problem of "who supervises whom" is separation of powers, not centralization of power. Project "alliances" lead to centralization of power, while the realization of separation of powers depends on institutions and culture - in the blockchain world, this institution and culture represents the "consensus standard."
