Analysis of Ethereum Improvement Proposal EIP-1559
If the Ethereum improvement proposal (EIP-1559²) is implemented, it will be the biggest change in the way users bid for block space on any major blockchain. While the proposed rules are relatively simple, the proposal will bring significant changes to users, miners, and wallet providers, and even affect Ethereum security in general.
Design goals
Design goals
EIP-1559 has four design goals:
1 Improve user experience
First price auctions, as used by blockchain platforms such as Bitcoin and Ethereum, are easy to understand and implement, but inefficient (see ref ³ and ref ⁴ for more on this). One of the biggest problems is cost estimation. EIP-1559 attempts to address this by enforcing the same fee rate for all transactions as much as possible. In most cases, users only need to decide whether to pay the fee, not how much to bid. This approach has the potential to reduce processing fees due to more accurate fee estimates.
2 Implement the Slack mechanism
The demand for block space usage may not be stable. As a result, some blocks are not yet more than half-used, while others are already full. The slack mechanism will allow certain blocks to grow as long as other blocks are relatively small. This allows a longer-term average block size limit to be enforced, while allowing for size variance between individual blocks.
3 Improve security
Blockchain platforms that rely solely on transaction fees for security (such as Bitcoin in the future) may experience problems when the block subsidy runs out⁵. EIP-1559 attempts to separate transaction fees from network security. Transaction fees will be destroyed in favor of a more reliable perpetual block subsidy to incentivize miners.
4 Prevent economic abstraction (economic abstraction)
secondary title
How EIP-1559 works: basic fee + tip
According to the agreement, users need to pay a minimum fee called BASEFEE, which is the first constituent element of the proposal.
Minimum fees are often not enforceable because the protocol cannot prevent transactions from being packaged via external price discovery. The agreement can always specify a price, but if the miner and the user agree on a lower price, the user can pay the miner according to the agreement, and the miner can refund the user outside the agreement.
EIP-1559 solves the above problems by completely destroying the base fee system, preventing chargebacks from happening.
Because miners receive no fees, users must incentivize miner transactions in other ways. Then from the user's point of view, the user must set two values when submitting the transaction:
First, the user needs to determine an additional fee called GAS_PREMIUM (hereinafter referred to as "tip") as the price for packing the transaction into the block. Typically, it has to be high enough to compensate the miner for a marginal increase in uncle block risk (eg: 1 Gwei). It allows raw first-price auctions between traders when the network is congested.
Second, users also need to determine a price ceiling (FEECAP), which represents the maximum amount they are willing to pay (including tips) for the transaction to be included in the block. This is necessary because in practice the base fee will go up and down every minute, and users with insufficient fee caps should have a chance to be included in subsequent blocks.
secondary title
Elastic block size cap
Block size flexibility is naturally as high as possible, but if the block size is not controlled, it will cause known incentive problems. The ability to increase uncle rates with larger blocks encourages miner centralization. In addition, verification costs must be controlled to maintain the trustless nature of the network⁶.
The elastic block size mechanism came into being. The goal is to allow miners to produce larger blocks at only provable cost. The increase in orphan rate itself is not a provable cost, because the orphan rate can be reduced by increasing the degree of centralization, but the destruction of handling fees in the protocol is a provable cost.
In EIP-1559, miners can occasionally exceed the block size cap to meet sudden demand, but only when there is genuine user demand. The original base fee prevents miners from augmenting their own blocks with spam transactions by introducing a real cost (the cost of burning).
EIP-1559 is possible because it replaces the existing hard cap on block size with two values: a long-term goal of 10m gas/block and a new hard cap of 20m gas/blockⁱ upper limit. (Please refer to Note i for related supplementary instructions)
Over the long term, the network will adjust the base fee up or down to achieve a desired average block size. When the block is smaller than the target size, the fee will gradually decrease to increase demand. When the block exceeds the limit, the fee will be gradually increased to reduce demand. Fees vary by a "distance" of 10m, but no more than 12.5% up or down per block.
secondary title
expected result
1 Improve user experience
To evaluate the user experience of users and wallet providers, we must consider the situation of the system under different congestion states.
No congestion status:
Whenever a block falls below the 20m gas max cap, users have no reason to tip beyond the minimum. This is true even if the base fee adjusts up or down over time.
In this way, packaging into blocks depends entirely on the base fee.
Users who are willing to pay both the basic fee and the minimum tip will be packaged into the next blockⁱⁱ (see note ii for related supplementary instructions). As long as there is no congestion, users can purchase block space at a fixed price. Just like when you go to Amazon today, you will find that your favorite products are marked with fixed prices, which is different from the bidding system in auctions. After seeing the price tag, the user either accepts the transaction or rejects it. This makes fee estimates for users and wallets highly predictable.
Users can even set the upper limit of the fee to a certain amount lower than the current base fee, so that they can be included in the block when the fee is reduced.
Occasional congested state:
When miners mine blocks above 10M gas, the base fee starts to rise. In fact, it keeps going up until a block with 10M gas or less is mined. If the capacity of the next block is 10M, the basic fee will be stable at the current fee position. If the next block capacity is less than 10M, the base fee will start to drop.
It's important to understand this. If the block size exceeds 10m for a while, transactions can become very expensive, eventually driving demand down. How fast are transaction costs increasing?
Assume that the basic fee in the 0th block is 1 billion wei/gas. And assuming the price of ETH is $240, the cost of a typical 21k gas transaction is $0.0005. After only 10 blocks of 20M gas, the transaction cost increases to $0.02. After 100 blocks, the transaction cost will increase to $657. This is the power of exponential growth.
When the block is close to the upper limit of 20M gas, most users with urgent transactions will participate in the tip auction, because the protocol no longer has a binary way to determine the priority of transactions (before, users were either willing to pay the base fee, or unwilling).
In this way, when the network is congested, the protocol will fall back to the existing first-price auction mechanism.
Even in tip auctions, traders can predict how high or low the base fee will be. If the base fee has a virtual starting value of 100 at block 0, it can reach up to 100*112.5% at T1, 100*112.5%^2 at T2, and 100*112.5%^3 at T3 ,So on and so forth. The same is true for fee reductions.
Taking the base fee with an initial value of 100 as an example, the blue area in Figure 2 shows the potential level it may reach in future blocks.
A state of increased congestion:
As can be seen from the above, EIP-1559 allows larger blocks to be used for a short period of time, rather than a longer period of time. After 30 minutes of a demand spike, the base fee for a simple transaction of 21,000 gas will exceed $1,000 (assuming base fee initial value = 1 Gwei).
The only way to get the base fee back to a more "normal" level is to develop blocks under 10m.
Just imagine if there were 100 blocks of 20m gas in a row. Continuing the previous example, the average cost of a transaction increases to $657. In order for the base fee to return to its initial value, 89 empty blocks must be developed. Alternatively, we could also develop 183 blocks of 5m gas, or 371 blocks of 7.5m gas, and so on.
Therefore, a typical pattern in high congestion situations is a series of very small or relatively small blocks followed by a series of large blocks. This also makes sense since users with high time preference can now increase their transaction/mining speed. But in the end, the base fee increased more than traders wanted, and they had to wait for the base fee to drop again.
Looking at it another way, slack brings many blocks' worth of capacity from the near future into the present. But it cannot generate more capacity by itself, and in the end, what is borrowed always needs to be returned.
2 Implementing Slack Mechanisms
By analyzing the way EIP-1559 deals with congestion, we can find that its slack mechanism is closely related to the maximum amount of change in the base fee between blocks. The slower it adjusts, the better the system can handle changes in trading needs.
A common theme of this variation is the periodicity of the day-night cycle or weekday-weekend cycle. In its current form, EIP-1559 does not allow blocks to grow overnight and then grow larger during the day, as these time periods are too long, causing daytime fees to skyrocket into the billions, and correspondingly Dropped rapidly to close to zero in the evening.
Thus, the relaxation mechanism works in the range of a few minutes to half an hour, beyond which it is ineffective. The longer the demand pulls the block size in one direction (for example: full block), the more intense the reverse pull effect of the increase in the basic fee will be.
3 Improve security
Many blockchain platforms have a deflationary monetary policy, where the number of new coins they issue gradually decreases over time. When circulation drops low enough, security becomes paid for by transaction fees. EIP-1559 abandons this model of maintaining security only by fees, because most of the transaction fees will not incentivize miners, but will be destroyed.
Therefore, it can be said that EIP-1559 provides a permanent block subsidy for miners to ensure the security of the protocol. However, it is more accurate to say that it optimizes the design options of the permanent block subsidy more appropriately. The reason is that the burning of fees has a deflationary effect on the currency supply, allowing new coins to be issued elsewhere without increasing the total supply.
Since a protocol with a permanent block subsidy brings miners a more stable source of income, it can be argued that EIP-1559 has a positive impact on the long-term security and stability of Ethereum.
4 Prevent economic abstraction
Before EIP-1559, strictly speaking, transaction fees did not need to be paid in ETH. While the network only supports fees paid in ETH, users could theoretically pay miners out-of-band in any currency type. However, miners can also get paid indirectly⁷ via MEV. For example, a miner can include transactions on a decentralized exchange (DEX) without charging fees because he can make money elsewhere through illegal front-running.
EIP-1559 largely addresses this issue. The base fee portion of every transaction is calculated in ETH and is always burned. In this case, no matter who pays, the ETH will disappear from the currency's supply. A miner is still free to choose a transaction with a lower base fee, but only if he has to pay the difference to the block subsidy in ETH. Miners will earn less ETH in the block, but users will keep more ETH. So it's a wash for the token supply, someone always has to pay transaction fees in ETH.
secondary title
potential problem
A major concern with EIP-1559 is whether miners can and will manipulate the base fee.
When the base fee is zero, miners receive the entire bid from users because there is no burn. Also keep in mind that tip auctions will only start when block space is in short supply.
Once fees approach zero, miners can simply keep fees at that level forever. If they never mine blocks above 10m gas, the base fee will never increase. If the demand never exceeds 10m (or whatever cap the miner decides to set below 10m), then the miner gets the full fee.
However, what is optimal for miners as a whole may not be the same for individual miners. This is the so-called collective action problem. If the upper limit of the block is 10m gas, and some users need to conduct transactions in the block of 20m gas, then only one miner can break this alliance and include the transaction of paying the tip.
For the basic fee system to stabilize and occupy a monopoly position, a miner is required to activate a soft fork (MASF). In MASF, more than 50% hashrate will guarantee to ignore any block larger than 10M gas, thus incentivizing that small group of people to follow the rules of this new implementation.
summary
summary
From the above analysis, we can see that EIP-1559 deserves its name to a large extent. It should increase the predictability of fee estimates, except for very short periods of high congestion, when the system falls back to the established first-price auction model. These periods are bound to last for minutes as the base fee increases exponentially, limiting demand.
This is possible thanks to the interesting building block of the base fee. The proposal opens up new design spaces by setting a minimum fee in the protocol: from elastic block sizes to permanent block subsidies, to greater resistance to economic abstraction, and better auction models.
The future looks bright for EIP-1559, but that doesn't mean it's already at its best. We only assessed how well the existing parameters matched their own design goals. Further studies of different configurations of this mechanism are warranted.
