Vitalik Buterin: What is the significance of EIP-2929 increasing Gas overhead in Berlin upgrade?

EIP-2929 increases the Gas overhead of some opcodes in order to improve the resistance to DoS attacks. At the same time, it becomes safer to moderately increase the Gas Limit.

Original title: "Vitalik: In the upgrade of Berlin, what is the significance of EIP-2929 increasing the gas cost? "

Summary

Yes, certain opcode calls will require more gas overhead, but it is necessary to improve safety.

But on the other hand, an operation that should have been cheaper all along (store and then load the same slot) ends up being cheaper.

Another important benefit is that modestly increasing the gas limit is now safer. The main reason for increasing the gas limit now is more of a state size issue.

VitalHu Deploys Blockchain-Based Coronavirus Screening Tool: Jinse Finance reported that Canadian technology company VitalHub Corp announced on April 8 the first deployment of a blockchain-based coronavirus screening tool, DOCit. The tool will be used by a long-term care facility in the Waterloo region called Sunny Side Home. The tool is designed to deploy and configure any screening or repetitive task, thereby setting reporting standards for care providers, operations staff, and national, provincial or territorial health authorities. [2020/4/9]

The EIPs that caused changes in the gas overhead in the Berlin hard fork are: EIP 2565 (reduce the gas overhead of ModExp modulo and exponentiation), EIP 2929 (increase the gas overhead of SLOAD and CALL), and EIP 2718 and 2930 to support supported The increase in gas overhead affects the compatibility of contracts and paves the way for future upgrades.

Market | Activity Index EOS ranks first: According to blocktivity.info data, the activity index of EOS ranked first is 6,101,280, the activity index of the second-ranked STEEM is 1,122,576, and the activity index of the third-ranked BTS is 949,526. The activity index of the fourth-ranked ETH is 574,337, and the activity index of the fifth-ranked BTC is 183,026. The Acitivity index is the number of operations performed on the blockchain in the last 24 hours. [2018/8/26]

Currently, the most important of these EIPs is EIP-2929. Its core content-zce is: the gas overhead of SLOAD is increased from 800 to 2100, the gas overhead of CALL (including STATICCALL, DELEGATECALL and other opcodes) and external contract query (BALANCE, EXTCODESIZE, etc.) is increased from 700 to 2600, but in this case Only happens the first time an address and storage slot is accessed in a transaction.

Sound | BM's evaluation of Vitalik's new consensus algorithm: it is a formal description of non-BFT finality: According to IMEOS reports, Vitalik recently published an article on his blog called "Guidelines for a 99% Fault-Tolerant Consensus". The algorithm only requires 1% of nodes to be "honest". This means that, in theory, an attacker would need to control more than 99% of the blockchain nodes to carry out the attack. So no more 51% attacks.

Members of the EOS community asked BM about his views and opinions on this article, and BM replied: This article does not give the time for example when it will be implemented, but it is safe to assume that this new algorithm will be implemented after the POS-based consensus takes effect. Implemented...so equivalent to implementing after the heat death of the universe...

“But this algorithm is a formal description of the non-BFT finality of Steem and Bitshare, which is very interesting.” [2018/8/16]

The purpose of this is to further improve the resistance to DoS attacks: early research shows that the biggest DoS vulnerability in the Ethereum protocol is storage access, and it is possible to create some blocks for simple and repeated access to a large number of accounts, processing The time may be as long as 80 seconds. The solution is a simple quick fix: make these operations last longer (storage access requires disk access) to consume more gas, and eventually the DoS problem is weakened by a factor of about 3. In the meantime, the client team has done some excellent work - implementing disk storage caching, reducing the number of database queries required for storage loads, and plugging this hole in the longer term.

Jinse Finance live report on Vitalik Buterin, the founder of Ethereum: The more locations of verification nodes, the more computing power is required: Jinse Finance reported on the spot on June 3 that at today’s Ethereum Technology and Application Conference, Vitalik Buterin, the founder of Ethereum, Made a keynote speech entitled "Casper and the latest developments in sharding technology". V God said that the responsibilities of the verification node are: verify the main chain; verify 2 shards (in order to generate blocks, to confirm blocks/crosslinks); generate shard blocks; generate blocks of the main chain; confirm shard blocks and cross-linked. V God pointed out: The more ether you have, the more positions you have for verification nodes, and the more computing power you need. This mechanism supports the decentralization of the protocol, and it will be fair to those who have less ether. [2018/6/3]

The re-pricing of gas overhead combined with these client-side improvements makes existing blockchains more secure and makes providing gas limits much safer than they are today. After EIP-2929, the main reason for avoiding a large gas limit increase is actually no longer the DoS attack problem, but the increase of the state size (these two aspects affect each other: when the storage load becomes larger, the gas consumed by the storage load will be more ). Therefore, state expiration and statelessness will become very important, and can even be said to be the top priority after "merging".

The second biggest long-term benefit of this repricing is a theoretical up to 3x reduction in the witness size for stateless validation. For code reads we also need a repricing (something like: 500 gas per access to 31 bytes of code), but this could be implemented in the future.

Please note: this EIP only increases the gas overhead for the first visit of each transaction. Subsequent visits are actually cheaper (100 gas in any case). Also, calls to precompiled always only cost 100 gas, including the first call. Doing so has the following positive effects:

Any SSTORE after SLOAD (or SLOAD after SSTORE) will become cheaper in the same slot. This is because the first storage read or write and the gas fee for accessing the storage slot have been paid, so the storage slot has been "warmed up", and the first read and write will be cheaper; thus, we need The gas paid is not 800 + 5000, but 2100 + 2900 (approximately), a reduction of about 800 gas. This could make transfers of tokens using the ERC20 protocol standard cheaper.

Self-calling just got cheaper

Calls to precompilations became cheaper (this is especially valuable for low-gas precompilations that need to be called very many times, such as ECADD)

The first access rule is added because when an account or storage slot is accessed for the second time, the account or storage slot should already have cached content-zce in it, so no expensive disk access is required. Of course, only one witness entry is required for two accesses to the same value.

Tags：

FTX