header-langage
简体中文
繁體中文
English
Tiếng Việt
Scan to Download the APP
Partnered With:

V: absolutely against the base level price prediction machine, ethereum L1 level function to be clearly limited

Read this article in 9 Minutes
Who are you rooting for in this battle between the core characters of ethereum 2.0?

Original title: "Vitalik: absolutely opposed to base level price prediction machine, ethereum L1 level function to be clearly limited"

Source: babbitt


针对以太坊 2.0 研究者 Justin Drake 最新提出的基础层价格预言机提案,以太坊联合创始人 Vitalik Buterin 撰文表示坚决反对,并提出了六大反对理由,此外他还表示,以太坊生态系统得益于强大的应用层代币生态系统,而不是通过 L1 层垄断所有重要功能。这场以太坊 2.0 核心人物之间的 battle,你支持谁呢?



(left: Justin Drake; right: Vitalik Buterin) & cake; & have spent


The following proposal was made by Justin Drake:& have spent


Simply put, we recommend adding a simple feed service to the beacon chain to track a small number of key assets. This service allows the establishment of a fully decentralized oracle that generates a price per trace asset at each epoch cycle boundary (i.e. 6.4 minutes).


Thanks @albert, @benjaminion, @dankrad, @danrobinson, @dcinvestor, @djrtwo, Eric Conner, Evan Van Ness, @karl, @khovratovich, Mehdi Zerouali, @mkoeppelmann, @paulhauner, @protolambda, @robert, @ryanseanadams, @sassal, @scott_ lew_is, @vbuterin for feedback and discussion.


Construct adds the price_data: PriceData field to the BeaconBlockBody, where: & NBSP; & have spent


PRICES_PER_ASSET = 8

# Type aliases Price = uint64 AssetPrices = Vector[Price, PRICES_PER_ASSET]

class PriceData(Container): # Currencies in the IMF SDR basket usd: AssetPrices # 0.01 USD price denomination eur: AssetPrices # 0.01 EUR price denomination cny: AssetPrices # 0.01 CNY price denomination jpy: AssetPrices # 1.00 JPY price denomination gbp: AssetPrices # 0.01 GBP price denomination


From the ethereum 2.0 consensus point of view, price_data & NBSP; The object is processed similarly to graffiti string. That is, price_data can be arbitrary data whose value is ignored by the beacon chain state transfer function and the biasing selection rule.


The honest beacon block proposer must populate PriceData with the best possible price for each trace asset, and these prices must:


1. Baseline -- 1 ETH for tracking assets;

2. Snapshot -- with the most recent PRICES_PER_ASSET& NBSP; Snapshot of the epoch (in AssetPrices vector, increasing in age order);

3. Face value -- in the smallest ISO 4217 currency unit (for example, Price(21309) if 1 ETH is $213.09);

4. Rounding -- rounding with half rounding (for example, how the 1 ETH is 22,732.66 yen, set to Price(22733));

5. Overflow -- if uint64 overflow, set Price(2** 64-1);

6. Underflow -- when the Price is negative, it is set as Price(0);

Example: honest majority eth-usd predictor we can build a mostly honest eth-usd predictor on each epoch cycle boundary.


Consider a smart contract that takes the median value of all prices corresponding to a given epoch (up to SLOTS_PER_epoch*prices_PER_ASSET=256& anywhere; Price point).


If more than half of the corresponding beacon blocks are generated by an honest verifier, the intermediate price is safe and cannot be manipulated by a dishonest verifier.


The basic characteristics of


The following are the basic functions of the service:


1. Consensus minimalism -- minimizes the impact on beacon chain state transfer functions and bidirectional selection rules by equating price_data with graffiti.


2. Subsidy delivery -- each beacon block contains a dedicated PriceData object to bypass the application layer block space restrictions and gas market;


3. Original price -- the feed price is a low level service about the original price, which can be mixed and matched with other price sources to create a mixed predictor with custom logic;


4. Bandwidth efficiency -- each tracked asset, 320 bytes of overhead for each BeaconBlock, serialized as type_byte_length(uint64) * PRICES_PER_ASSET = 64& NBSP; Bytes. The bandwidth overhead of block propagation is small, especially in the case of network compression.


5. Can be used by other chains -- this service is available not only for ethereum 1.0 and ethereum 2.0, but also for other blockchains. It can also be used off-chain, for example to estimate gas costs in fiat currency through a fully decentralized wallet.


6. Initial deployment -- this feed service is an optional ethereum 2.0 infrastructure that can be deployed when it is ready. The DeFi ecosystem of ethereum 1.0 will benefit, as in phase 1. & have spent & have spent


Hello price management


Tracking the maintenance of a set of assets is a governance issue, and we recommend following the following community norms:


1. Minimum feasibility -- the tracking asset set should be maintained at the minimum feasible level, i.e. only assets essential to ethereum's long-term health and success should be included. The five initial assets (dollar, euro, RMB, yen, pound) reflect three key principles :(1) credible neutrality; 2. High liquidity; 3. Wide application)


2. Crude consensus -- new assets (such as gold, oil, bitcoin) can be added to the tracking set through a hard fork driven by crude consensus.


3. Backward compatibility -- backward compatibility after adding an asset cannot be removed from the trace set to maintain backward compatibility. & have spent & have spent


Verify the runner behavior


We discuss the expected behavior of various types of validators:


1. Honest operator -- by definition, an honest operator runs validation nodes that conform to agreed rules and reports correct price data.


Lazy runner - a lazy runner (such as a busy amateur pledger) whose laziest behavior is to run the default verifier client Settings. Therefore, a lazy runner should run the verification node that reports the correct price data. (since the computing overhead for loans, CPU time, RAM, disks, and so on is small, there is no need to consider computing latency.)


3. Rational operators -- rational operators seek to maximize revenue and will consider price data manipulation to achieve verified maximization of extractable value (VEV, similar to the concept of miner extractable value). However, because the PriceData object is public and encrypted signed, dishonest validators can be identified by the community. Potential penalties include :(1) loss of reputation; 2. The verification balance is confiscated; Validator client guide validating clients must be able to query external sources to derive prices on the epoch boundary. The following is the client implementation guide:


4. Variability -- various price sources must be consulted for robustness, including on chain sources (e.g. Uniswap V2, MakerDAO, Chainlink, see oracles. Club) and off chain sources (e.g. Coinbase, CoinMarketCap);


5. Security -- the attack vectors generated when querying untrusted apis must be handled carefully;


6. Be aware of external issues -- the source of anomalies must be adequately addressed (e.g., using fitness tests and weighted medians);


7. Customizability -- the client runner should be able to easily change the default price derivation logic and customize its price source;


8. De-correlation -- the realization shall strive to maximize the price inference logic to de-correlate (e.g., across implementations, versions, slots);


DApp guide


1. History accumulator -- each PriceData object can be validated against historical_roots in BeaconState. In order to improve gas efficiency, the SHA256 Merkle verification data (witnesses) can be effectively compressed into a SNARK (optimized by plookup and other technologies);


2. Risk analysis -- dApp must carefully evaluate the price manipulation risk of the dishonest verifier and make an informed security assumption (such as the verifier's honesty assumption);


3. Hybrid forecaster -- dAppp should consider combining important feed services with other price sources. Feed-in service is an infrastructure designed to increase the selection and robustness of ethereum ecosystems; & have spent & have spent


Below, battle from ethereum co-founder Vitalik buterin: & NBSP; & have spent




"I absolutely disagree!First, this is a fundamental change in the nature of blockchain technology. We now have a feature where the correctness of blockchain progress can be fully verified programmatically. Availability (that is, non-censorship) can be verified by online nodes, and there are even techniques for low-latency online nodes to reach a consensus on whether the blockchain should censor transactions.


On the other hand, the purpose of this proposal is to introduce a chain property that, even in principle, cannot be programmatically verified under any assumptions. Even in a future world, there is no obvious consensus on the correct value of inputs (for example, there was a civil war in one of these countries, where each side claimed its currency was the "real" USD/CNY/JPY).


Second, the proposal relies on the honest majority, but a lot of what we've done with ethereum 2.0 is basically to get rid of the honest majority assumption and try to create a "second line of defense" where the honest majority fails,For example: hosting proof, an attempt to change the security assumption of aggregation from "must be honest" to also cover participants who are not evil but lazy; Proof of data availability and fraud proof, allowing 51% of the attack chain to reject invalid or unavailable blocks; Users can run full nodes, not just light nodes; The review and testing techniques mentioned above; No activity leakage mechanism, and its in > Restore when 1/3 offline attack occurs; Making a design that relies on an honest majority is the opposite of all these advances.


Third, it undermines the neutrality of the agreementAnd opens the way for further neutral compromise to slip into the abyss. The proposal (1) elevates "defi" to the privileged application category and (ii) elevates a specific set of asset/price indices. It also introduces the political risks of underlying governance, such as having to decide which currencies are important enough, which application classes are important enough, how to judge emergencies, and so on.


Fourth, it closed the door to innovation in the design of predictive machines, for example, a natural alternative to this design is that the price at time T should only be agreed at time T+1 day to provide room for attacks on the chain, for exchanges to stop working for a longer period of time, and for unexpected failures of the functional API in general. There are many ways to design a prognostic machine, and the L1 approach to ecosystem governance does not seem to be the right answer.


Fifth, it increases the risk of the centralization of the verifier of stakingBecause clients need more automatic updates to maintain their machines, this increases the risk that validators will blindly follow the instructions of the client developers (or people will abandon being the validation node altogether and switch to the validation pool).


Sixth, it doesn't actually provide more security than application-layer toker-based predictors (such as Augur, etc.).The market value of MKR is about 2 million ETH, so it is obvious that the application tokens can obtain significant market value. We expect the initial staking of ethereum 2.0 to be in the range of about 2 million ETH (possibly 10 million ETH in the long run). Therefore, the security level of the base layer predictor is not much higher than that of the application layer predictor, which seems to be an order of magnitude difference at best.


I actually think that we should move in the opposite direction, explicitly limiting the functionality at the bottom so that there's room for an application-level ecosystem to develop other tools at the bottom. Augur has been running well, other predictors (UMA, MKR, Chainlink...) It also exists.


The einsteum ecosystem benefits from a powerful application-level token ecosystem, rather than the L1 /ETH monopoly on all important functions. This is because the etheric fang ecological system there is a great demand for public goods, and ready to provide these public products supply of etheric fang co., LTD. (the etheric fang foundation holds 590000 ETH, as well as some whales hold money), so modify the etheric fang agreement, to print more etheric fang used for these purposes, are politically difficult (for good reason). However, application-layer token can provide these common goods, such as Gnosis, which has done a lot of work with smart contract wallets, and which now maintains openesom, and so on. The application layer token can even directly use quadratic financing to finance public goods. Therefore, we should deliberately seek and design a symbiotic relationship with such application-layer tokens, rather than regard them as a testing ground that can be absorbed by the base layer. "


So far, most ethereum participants seem to support Vitalik's view. What do you think? & have spent & have spent


The original address: https://www.8btc.com/article/595058




Welcome to join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Official Twitter Account: https://twitter.com/BlockBeatsAsia

举报 Correction/Report
Choose Library
Add Library
Cancel
Finish
Add Library
Visible to myself only
Public
Save
Correction/Report
Submit