Ethereum London upgrade is coming, what do miners and users need to pay attention to?

Original title: "Here comes Ethereum "London" upgrade"
Written by: ETH Chinese website

The eleventh hard fork "London" upgrade of the Ethereum network will be activated at block height 12,965,000 (estimated to be carried out on the evening of August 5, 2021, Beijing time). This upgrade mainly includes 5 EIPs, namely EIP-1559, EIP-3198, EIP-3529, EIP-3541 and EIP-3554.

What do I need to prepare?

Node operators need to upgrade the client version they are running. The following are the client versions that support the "London" mainnet upgrade:

• go-ethereum (geth): 1.10.6

• Nethermind: 1.10.79

• Erigon (formerly TurboGeth): 2021.07.04-alpha

• Besu: 21.7.1

• OpenEthereum (formerly Parity): v3.3.0-rc.4

• EthereumJS VM: v5.5.0

Tim Beiko (Ethereum ACD Coordinator/EIP-1559 Advocate) has recently received a lot of information about London upgrade mining, among which there are many misconceptions. He gave a detailed explanation of the work, precautions, and related resources that miners need to do in the London upgrade, and also popularized how users should set tips in an environment with MEV:

Miners' Notes

When the London upgrade is activated, the gas limit of the forked block will become twice that of the last pre-fork block, which can be seen on the forked testnet.

For example, the gas limit of the block before the Goelie test network fork is about 8m, and the gas limit of the forked block is about 16m.

However, if they want to keep the throughput on the chain constant, miners need to keep the new gas limit constant (the average block capacity is expected to be about 50% full).

If miners do nothing, the block capacity will continue to be capped at 15m gas, and the block size will slowly decrease. To avoid this, miners need to raise the gas limit to 30m after the fork (if they want to keep the average block capacity at 15m gas).

Miners do not need to restart your mining nodes. Each client that supports the London upgrade will implement this through JSON RPC. Please see the API call of each client in the London upgrade announcement:

• go-ethereum (geth): miner_setGasLimit

• OpenEthereum: parity_setGasCeilTarget

• Besu:miner_changeTargetGasLimit

Here is the link to the JSON RPC documentation:

• go_ethereum

• OpenEthereumOrg

• ConsenSysQuorum

Again, miners do not need to restart their nodes, they only need to set a new gas limit in the JSON RPC call.

User Notice

Ethereum users or ETH holders do not need to take any additional actions during the upgrade, but they should pay attention to the upgrade notifications issued by wallets or other service providers. In addition, Eth2 validators should also ensure that your PoW (eth1) nodes are upgraded to the latest version before the London upgrade.

The second big question is about the setting of priority fee. In short, the tip is paid to the miner to handle the additional risk of uncle blocks appearing when packaging your transaction. Bigger blocks -> slower broadcasting on the network -> higher uncle block risk.

There is a way to visualize this situation: increase the gas limit of the previous block from 12.5m to 15m. The left point shows the probability of uncle blocks appearing in the range of 12.5, and the right point shows the increase in uncle block rate when we adjust it to 15m: the additional 2.5m gas brings.

Therefore, if a user wants to pay a high enough tip to miners to get them to include your transaction, then the tip needs to be a net benefit to them on average, and this is related to the probability that the user's additional transaction will be included in the uncle block.

In the pre-MEV world, this calculation is easy because the block reward is separate from the uncle reward, so we can just use a fixed value (0.5-1 gwei).

The challenge with MEV now is that when a block becomes an uncle block, the miner will lose the income in the MEV transaction bundle and part of the block reward, and the MEV income will vary greatly!

Therefore, the best approach is to look at the median or 80th percentile value of MEV transaction bundles and compensate miners for the risk of losing that revenue. Flashbots has some dashboards that show this data:

https://dashboard.flashbots.net/miners

The reason for choosing the first 50/80% of block revenue as a tip is that it doesn't make sense for the average user to try to compete with a 10ETH+MEV transaction bundle. This situation is rare, and if it does happen, the average user should not mind waiting for the next block.

So how do you choose a tip value? @barnabemonnot did a lot of math on this and gave a simple strategy.

So, looking at the Flashbots dashboard and the chart above, we can see that 2-3 gwei is enough to cover the opportunity cost of 80-90% of MEV transaction bundles, so this would be a good default value.

Right now this isn't super intuitive and the MEV calculation makes it hard to set it once and forget it, but hopefully in a few weeks APICs like @ETHGasStation and @gasnow_org will abstract this data away.

References:
https://twitter.com/TimBeiko/status/1422960629921550353?s=20

What does the London update include?

EIP-1559: Eth1.0 Fee Market Changes

EIP-1559 was proposed in April 2019. It is one of the most widely discussed and anticipated changes in Ethereum's history. It is also the most far-reaching measure in the "London" upgrade. It aims to reform Ethereum's current "First Price Auction" fee market. The core goal is to improve the user's trading experience.

After the implementation of EIP-1559, Ethereum's transaction fees will be transformed into a base fee plus a tip (BaseFee and PriorityFee respectively), where the base fee will be automatically destroyed instead of being distributed to miners (miners' income is part of the tip), which is widely believed to have a profound impact on the Ethereum network's economic model. One of the determining factors of Basefee is the saturation of the previous block, which means that applications and users can more easily predict the cost of transactions, and after paying the base fee, transaction confirmation is also guaranteed.

EIP-1559 also allows for flexible changes in block space, i.e., elastic block space (currently allowing 200% block capacity), and the base fee will also be adjusted accordingly. That is, the current gas limit of Ethereum blocks is 15m, and after 1559, it will be used as gasTarget (gas target), and gasLimit will reach 30m. If the gas used by the block is higher than the gas target, the base fee will increase until the network demand drops to the target again, otherwise the base fee will be reduced. In addition to the base fee, there is another factor that affects the transaction fee, namely the tip, and the amount of the tip is not limited by the protocol. In general, the tip is used at a lower amount to pay for the uncle block risk. During peak trading periods, transactions return to the first price auction mechanism through tips.

In addition, EIP-1559 adds a new transaction type where users can specify the maximum amount they are willing to pay. When they send this maximum fee to the miner, they will receive a refund of the difference between the maximum fee and the sum of the base fee and the miner's tip. If the amount of ETH on the sending transaction account is less than this maximum fee, the transaction cannot be executed.

The following link shows the burning status of the basic fee:

https://watchtheburn.com/

https://www.etherchain.org/burn

EIP-3198: BASEFEE opcodes

This EIP is paired with EIP-1559. It simply adds a BASEFEE opcode that returns the base fee of the block the transaction was executed in. This will allow smart contracts to access this value on-chain, facilitating the submission of fraud proofs and the creation of trustless gas price derivatives.

EIP-3529: Reduce gas refunds

Another major change introduced in London was the removal of gas rebates for the SELFDESTRUCT opcode and the reduction of gas rebates for the SSTORE opcode. While the original intention of the rebates was to incentivize developers to clear state when possible, the reality is that this led to the emergence of Gas Tokens, which increased state size. With this rebate, Gas Tokens can fill up the state when gas prices are low, and then get rebates for executing these transactions when gas prices rise.

Among other things, gas refunds can lead to changes in block execution time. Before the London upgrade, up to 50% of the refunded gas could be used to perform further calculations in the same block. That is, in practice, the maximum block capacity could be up to 1.5 times the gas limit. EIP-2539 reduces "execution gas refund" from 50% to a maximum of 20%. This change will help offset the additional block size changes introduced by EIP-1559, as EIP-1559 allows blocks to use twice the gas limit.

EIP-3541: Reject new addresses starting with 0xEF bytes

EIP-3541 is a simple change that lays the foundation for more extensive EVM improvements in the future, see EIP-3540. This EIP will make it impossible to deploy new contracts that start with 0xEF bits. Existing contracts will not be affected.

EIP-3554: Difficulty bomb delayed until December 1, 2021

EIP-3554 Delayed Difficulty Bomb, also known as the Ice Age. The Difficulty Bomb or Ice Age is a mechanism introduced by Ethereum to "freeze" mining when the network transitions to Proof of Stake. This has been done three times in the past, with Metropolis (EIP-649), Constantinople (EIP-1234), and Muir Glacier (EIP-2384). Although the previous delays were quite long, this time the core developers have opted for a shorter delay, pushing the difficulty bomb to December 1, 2021.

Where will London go after this?

Ethereum Upgrade Path @trent.eth

Following the “London” upgrade, Ethereum will also usher in the Altair upgrade for the beacon chain and the merger of Eth1 and Eth2 (The Merge).

The main contents of the Altair upgrade include changes to Ethereum 2.0 staking parameters and reward and punishment calculations, as well as the introduction of a synchronization committee for the implementation of light clients. It is currently in the testnet stage.

After this upgrade, the related work and progress of The Merge will become the focus of the entire community. In December 2020, the Ethereum beacon chain was created and has been running for nine months, but the current Ethereum blockchain and the beacon chain are still two parallel Ethereum blockchains. The specific content of the merger of Eth1 and Eth2 is to use Eth1 as the execution chain, and its status, execution, transactions, etc. will be incorporated into the beacon chain as the consensus chain. Each PoS beacon block will contain the data of the execution layer. At the same time, the Ethereum blockchain will no longer use PoW consensus when creating new blocks, thereby realizing the consensus upgrade from proof of work (PoW) to proof of stake (PoS).

On July 22, researcher Mikhail Kalinin, Ethereum core developer Danny Ryan, and Ethereum founder Vitalik Buterin jointly released EIP-3675, formally proposing the merger of Eth1 and Eth2 as an improvement proposal and introducing this PoS consensus upgrade.

Regarding the development path after the "merger", Vitalik explained in his recent ETHGlobal speech, including long-term plans such as sharding, data availability sampling, and stateless Ethereum. See "How will Ethereum develop after the merger of Eth1 and Eth2?"