Did you know that Ethereum EIP-1559 may not be stable?

Although EIP-1559 is intended to benefit the Ethereum community, it is actually unstable. Since Ethereum has gradually been coveted by other public chains for its high gas fee, the launch of EIP-1559 has attracted much attention and high hopes. It is understood that since the launch of EIP-1559, the net issuance of ETH has decreased by about 68%, equivalent to 1.48 million ETH being destroyed. This is also hailed as the most bullish catalyst in Ethereum's history. But at the same time, some remarks about the defects of EIP-1559 have begun to circulate, and some even say that the existence of EIP-1559 does more harm than good. In fact, EIP-1559 does not solve the problem of high gas fees. From an economic point of view, the level of gas fees depends on the market supply and demand relationship, so the fundamental way to reduce handling fees is to expand capacity. But in addition to this, what are the shortcomings of EIP-1559? In this article, we will conduct an in-depth discussion on some of the disadvantages of EIP-1559.

In this article we will use the base gas price to reflect the performance of EIP-1559. With EIP-1559, transactions have a new field and are serialized using a new format. Instead of specifying a unique gas price, the fee is specified as the maximum gas price to be paid and a miner tip (called a "precedence fee"), which determines the amount that will be paid to the miner above the base gas price. Since the tip must be a positive number, the gas price to be paid will always be higher than or equal to the base gas price. When a transaction is processed, the total fee (price multiplied by the gas consumed) is divided into two parts: the base amount is burned, and the miner tip amount is paid to the miner of the block.

When most blockchain protocols need to destroy native coins, it is generally to punish a party for misbehavior. If the misbehaving party is rewarded, then the misbehaving party can condemn themselves to avoid being condemned by others. In these protocols, no matter who condemns, the series of actions performed by the parties always results in the coins being burned. However, in the case of EIP-1559, the same actions that caused the tokens to be destroyed can restore the tokens if they occur in a different order or time. Similar to MEV, if all parties involved can cooperate, the tokens can share the revenue.

The following are the short-term incentives for protocol participants:

• Miners hope to collect more fees from users or by lowering the base gas price.

• Active users would like to pay less, either by lowering their tips or by lowering the base gas price.

• Passive users want the base gas price to be high so that more tokens are destroyed and the value of the token goes up.

All active parties will benefit from a reduction in the base gas price. The main problem is that joint action between active users and miners to reduce the base gas price is difficult to achieve. These are the obstacles to coordination:

(1) Both groups are anonymous and dynamic,

(2) The number of active users is high,

(3) coordination requires some initial time or money effort,

(4) Defectors may disrupt plans,

(5) Remaining users can take advantage of this situation by transacting with lower fees and raising the base gas price without participating in a joint effort.

While EIP.-1559 dismisses coordination, the EIP likely did not foresee the huge incentives that exist today to lower the base gas price. Today, priority tips make up just under 6% of the base gas price. In other words, miners can earn 16 times more from transaction fees by lowering the base fee. Eric Voskuil’s book on cryptoeconomics explains that transaction fees are the cost of censorship resistance, and Ethereum currently pays too little for it. This means that, at least in theory, all transactions can be easily suppressed. The suppression of transactions causes the base gas price to quickly drop to zero, increasing miner revenue after wallets adapt to the change.

If the base gas price drops to zero, both users and miners can split their savings 50/50 for future transactions. Users will trade at a 50% discount, but miners will receive 8x the revenue from transaction fees. Since transaction fees paid (burned + not burned) have reached parity with the block subsidy, this results in a 50% increase in paid block rewards. Since miners’ net income is only a small fraction of the reward due to electricity and other operating costs, this new revenue source could represent a 300% increase in net income, even if 50% of the new revenue is shared with users. This is obviously a very unstable situation. If only the Ethereum blockchain can coordinate these two groups, or if transactions can be censored, then EIP-1559 will be game over.

If the base gas price of a transaction magically drops to zero instantly through a perfect coordination system, more users will attempt to transact, pushing transaction fees back to the same level as before. However, assuming a coordination system between miners exists, the base gas price can remain at zero forever. Miners will have a stable net income that is 400% higher than before.

In this article, we showed several ways (some cheap) to coordinate lowering the base gas price. While some of these coordination mechanisms may be imperfect and may fail at first, it is generally said that vulnerabilities will only get worse. People will refine coordination systems to achieve their goals. The sole existence of so many ways to achieve coordination shows the current fragile state of the Ethereum network under EIP-1559.

Notably missing from the research presented in this article is an analysis of Ethereum wallet fee management code. Wallets are expected to correctly handle the reduction in the base gas price and revert to arbitrarily increasing the tip to confirm transactions, as pre-EIP-1559 wallets did. However, wallets may implement sanity checks to prevent the old fee market from operating again. A deeper analysis would require looking at the source code of most existing wallets to see if wallets compatible with EIP-1559 have hard-coded restrictions that prevent the payment of tips above the base gas price.

Equally important, the closer the PoS merger gets, the higher the incentives for PoW miners to act individually and collectively.

Miner coordination without a “classic” 51% attack

We assume that a majority of miners (say 60%, in terms of hashrate) are willing to cooperate to lower the base gas price. We call them the Coordinating Majority (CM). A CM has the ability to arbitrarily increase or decrease the block gas limit, and our first coordination method will exploit this ability.

EIP-1559 compares the existence of the CM group to a 51% attack. But this is not the case. The classic 51% attack by miners is to discard the blocks of a minority of miners to increase the profits of the majority of miners. In CM, the majority of miners provide benefits to the minority. From the perspective of miners, this is the opposite of an attack and a donation. From the perspective of miners' rationality, all miners are incentivized to cooperate with the CM behind the scenes, even if they do not join the CM for political reasons.

This brings us to the first way to circumvent the basic Gas price control mechanism. Currently the block Gas limit is 30M Gas. If 60% of miners create blocks that consume 15M Gas, and the rest of the miners fill them to 120M Gas, this equates to a throughput of 57M Gas/block. We can assume that the demand for Gas will satisfy this proposal, since the BSC chain has an 80M block Gas limit and will fill up some of its blocks. Therefore, if CM decides to increase the block Gas limit 4x to 120M, and CM artificially limits the Gas consumed by its own blocks to 15M, all miners win. This holds true even if the decrease in Gas price is inversely proportional to the increase in throughput.

To illustrate why this strategy benefits all miners, we can see that if a minority completely consumes all available Gas in a block, each non-CM block will increase the base Gas price by 12.5% ​​(the maximum increase allowed by the protocol). Each CM block with a target of 15 million Gas will cause the base Gas price to drop by 10.9%. Because CM mines 60% of the blocks, the result is a continuous net decrease in the base Gas price. It takes 190 blocks to reduce the base Gas price by 90%.

The attack requires a preparation phase, where the CM raises the Gas limit to 120M, which takes 1422 blocks (about 6 hours). In such a short period of time, the community will not have a chance to coordinate a response, let alone a hard fork. But what happens during the rapid expansion of the Gas supply is unclear. If demand does not match supply during the preparation phase, the base Gas price may fall. It is possible to make the base Gas price go to zero just by expanding the supply, and the tip will also decrease, resulting in a net reduction in miner rewards. The CM can think of it as an investment in future returns, but at today's Ethereum prices (6-hour revenue), the "investment" may reach $800,000. However, we will come up with a better coordination mechanism.

Miner coordination for a 51% attack

If we assume the block limit is raised to 120M Gas, but CM starts orphaning blocks from non-CM miners that consume more than 15M Gas (a 51% "attack"), then it would only take 24 blocks with a Gas price of 10% of its original value to reduce the base. To avoid an open orphan war with other miners, CM= would provide a one-line code patch (or even a configuration change) to all miners in advance so that miners would artificially cap their block limit to 15M.

The end result is that within 6 hours, the block Gas limit may be raised to 120M, the network capped at 15M Gas blocks, and all miners will collect more than 2 additional Ether in each block, doubling their income.

Third-party coordination

EIP-1559 aims to fill all blocks to about 50% except during periods of high demand. One of the features of EIP-1559 is that if all odd blocks are empty, and all even blocks are 100% filled with 30M of Gas (achieving the same average target throughput), then the base gas price continues to drop. After only 300 alternating empty and full blocks (about 1.25 hours), the base gas price reaches 10% of its original value. Note that it is not necessary for 100% of miners to decide to create empty odd blocks. If only 30% of miners create empty odd blocks, then after 300 blocks the base gas price drops by 50%, and again by 90% after 1000 blocks (about 4 hours), and we see that a majority of miners is not required.

Note that if all miners fill less than 47% of blocks, the base gas price will also drop and reach 10% after 316 blocks, but the throughput will drop slightly by 1.5%.

The breakeven point to compensate miners for any temporary losses from mining empty blocks is 1.25 hours earlier. With only 40% of miners participating, and in just 15 minutes (60 blocks), the base gas price can drop by 20%, which is enough to compensate them for any losses from empty blocks. The downward spiral in the base gas price has been triggered.

Now we will invite a third party to help reach the break-even point. Suppose there is a hidden coordinator named Charles who is willing to invest in rewarding miners who create empty blocks for 15 minutes, compensating everyone for their potential losses when creating an empty block, plus a 10% bonus. Assuming a block every 15 seconds, Charles' maximum cost corresponds to the tips that could have been paid in 60 blocks (30 tips to the empty block producer) multiplied by 1.1 to provide the bonus. We assume an ETH price of $3200/ETH, each block contains 2.12 ETH, and 2 ETH are destroyed in fees paid by users.

Assuming Charles uses the average tip over the past 15 minutes as a reference, this would result in a budget of $13,000. With only 40% participation, if the market responds to the cap on the gas supply by increasing the tip amount, then after just 15 minutes, all miners have earned a total of up to $40,000 in fees, preventing them from being destroyed. The 40% who participated earned an extra $16,000.

Charles can accept donations on-chain to create incentives for a continued even/odd mining model, but this is not required.

Charles can say that the bounty will only be paid if less than 40% of people participate in mining the empty odd block. This means that miners never lose money: either Charles compensates them, or they are compensated by users who tip at a higher gas price. Since the most likely outcome is miner participation, Charles can save the bounty and re-use it for the next round. After miners realize their ability to lower the base gas price, even if Charles leaves, miners will most likely just continue mining empty odd blocks as the new Schelling point! The downward spiral has begun.

Coordination through smart contracts

If you don’t believe that a new Schelling point can be reached, or that a third party can be trusted to coordinate it, then we show that the power of smart contracts can achieve the same goal: we can replace the external party Charles with an open and secure smart contract that brings miners together. Let’s call it the Gatherer. Miners will interact directly with the Gatherer, without trusting any external entity. As long as the miner understands the contract, any miner can trigger a chain reaction that increases the miner’s revenue.

It is obvious that it is possible to build a gatherer in Ethereum, using the BLOCKHASH opcode to get the hash of a block from the last 15 minutes or more. The gatherer must provide some incentive for users or miners to submit complete block header information to the gatherer contract, which will check if this information matches the securely obtained block hash. Through the block header, the contract can know all the information about its past blocks and discover which miners created empty odd blocks and compensate them accordingly. Detecting empty blocks is simple because the transaction root hash corresponds to an empty trie hash. After all, coordination is the benefit of Ethereum.

A gatherer can easily decide to pay the bounty or use it for another 15 minutes, so the most likely outcome is that the gatherer only needs an initial investment of $13k, which will last forever. If it takes more time to raise a tip due to a shorter gas supply, the bounty and bounty duration can be extended. This is a question of parameterization, but not feasibility.

Coordination of a small number of miners and private users

One of the desired properties of EIP-1559 is that users can better estimate transaction fees and avoid overpayment. But at the same time, the base gas price must quickly adapt to congestion. Currently, it is common for the base gas price to increase by 100% in a single day.

The huge fee variability creates an incentive for users to negotiate fixed rates with miners. As we will show, a mining pool we call Mina can offer users lower transaction prices by mining empty blocks while gaining a huge revenue boost. EIP-1559 raises this problem and ignores it: miners who do not mine empty blocks will undermine Mina's plans. However, our simulations show that this is not the case.

This is how Mina does business: Mina offers a fixed-rate contract to users, whom we call Mina’s “private clients.” The product is valid until a deadline given by a block number, and if the private client attempts to double-spend a given transaction before the deadline, they agree to be penalized.

The penalty mechanism is implemented by a smart contract we call Penelope. The collateral deposited by a private client is approximately equal to the cost of two transactions for Penelope. When a client sends Mina a private transaction T with nonce N, it also signs a message M asserting that he will refrain from creating any double spends T´ until a deadline D. The message to be signed is as follows: “nonce N will be reserved for transaction TxId(T) until block D”. Private transaction T will specify a maximum fee below (i.e. 3%) the current base gas price, which is agreed upon by the private client and Mina. If the private client double spends nonce N, then Mina will send the proof to Penelope and collect the bounty. Note that Mina’s private clients are more likely to be large cryptocurrency exchanges.

Mina is now free to collect as many other private transactions as she wants from other clients under the same protocol. Note that if Mina collects enough commitments (e.g., she can fill a full block), then this also means that she has removed the gas to be consumed from the market, and filled blocks will have fewer, and the base gas price will be lower, even before Mina mines an empty block.

Currently in Ethereum, the average tip is less than 6% of the base Gas Price. To simplify our explanation, let's assume the base Gas Price is 100 tokens and the average tip is 6 tokens. Mina's private clients' transactions pay her at most 103 tokens (about 3% less than 106), but they specify a tip of 15.5 tokens (much higher than the other 6 tokens). Mina cannot abuse these private transactions because they are invalid in a block with a base Gas Price of 100. Mina will wait. Let's assume that all blocks are half full and that the base Gas Price is stable. Let's assume that the Gas Limit of a block is 30 units of Gas. Now we explain how Mina will benefit from these private contracts by mining empty blocks.

Once Mina collects a block full of transactions, she starts mining empty blocks. For each empty block she mines, the base Gas price decreases by 12.5%, and the base Gas price becomes 87.5 coins. If she happens to mine a block after mining an empty block, she can fill it up with private transactions and earn 103–87.5=15.5 tokens for each Gas unit used. Since she fills the block completely to 30M Gas, she earns 15.5*30=465 tokens. Every time she mines an empty block, she loses about 6*15=90 tokens in revenue. This means that as long as Mina can create two blocks in a row every 5 blocks, she can get a positive return.

A miner with 45% of the hashrate has a higher chance of mining two consecutive blocks, rather than one every 5 blocks. Therefore, it is wrong to assume that the majority of miners need to profit from a reduction in the base gas price, as we show that 45% of miners can actually profit.

We take into account here that after an empty block is mined, the remaining miners will fill their blocks to consume all the Gas that was not consumed before, but this may not be the case. If not all Gas is consumed, then Mina's strategy may trigger a continuous decrease in the base Gas price while the overall transaction cost remains the same to the benefit of all miners. In practice, what actually happens depends on how wallets are programmed to increase tips in case of confirmation delays.

We have not yet evaluated the impact of Mina’s strategy on the remaining users (those who have not entered into a private contract with Mina). Knowing Mina’s strategy, users willing to make a transaction might decide to wait until a block with a lower base gas price is created, and then broadcast their transaction immediately to compete for block space with a low base gas price, setting a lower maximum fee. Using the numbers from our previous example, one possibility is a tip of 10 tokens with a maximum fee of 100 tokens, successfully saving 6 tokens. This strategy benefits all miners, including Mina. After the empty block, the miners’ income almost doubles. Therefore, their incentive remains to protect Mina, even if they are not part of Mina.

Autonomous Mining and EIP-1559

A miner who wants to presell block space but doesn't want free riders following his empty block might try autonomous mining, attempting to create two private blocks in succession and publish them together. If that's not possible, then the first block will be released and included later as an uncle, losing at least 1/8 of the block subsidy. Currently, if 80% of fees were shifted from base gas price to priority tips, then fees would provide a higher income than the block subsidy, meaning autonomous mining could become a rational strategy for miners with a large percentage of hashrate. Likewise, once a drop in base gas price is triggered, it could end up with zero base gas price.

After EIP-1559 “Merges” PoS

One of the problems with EIP-1559 in Ethereum PoS consensus is that a miner knows in advance when he will be able to mine two blocks in a row. A miner can pre-sell block space at a price lower than the average Gas price, and when it is his turn to mine two consecutive blocks, mine an empty block and then mine a private filled block. Due to the openness of leader selection, Ethereum PoS also makes it easy for miners to coordinate actions with previous or subsequent miners to reduce the base Gas price.

Solution

We cannot find a perfect solution to the incentive problem of EIP-1559. The solution we propose is to only destroy a portion of the base fee (i.e. 20%) and give the rest to miners, which is preferable to mining pools.

Some changes could reduce the incentive to coordinate. The simplest patch would be to reduce the rate of change of the base gas price from 12.5%/block to around 3%/block, but this would change the properties of EIP-1599 to communicate congestion to users. Still, there is no way to prevent long coordination via smart contracts. Another partial solution would be to set a hard limit in block size to 30M (several EIPs attempt to do this).

Since EIP-1559 has reduced the cost of transaction censorship below acceptable levels, it cannot be saved without reducing the fees consumed. Time will tell whether EIP-1559 is stable or becomes unstable without the proposed changes. Community pressure on miners can mitigate the risk, even if it is unreasonable for them not to bypass EIP-1559.

Summarize

While the intention of EIP-1559 is beneficial to the Ethereum community, especially passive Ethereum holders, we believe it is destabilizing. When all active participants can highly benefit from coordination, and the cost of coordination is low, only a spark is needed to trigger a change that reverts EIP-1559 to the previous state. A chain of events can quickly converge to a new reciprocal Schelling point to eliminate the base gas price. The root of the problem is that the base gas price can be changed by transaction censorship, and the censorship cost of Ethereum has become very low. If the average throughput remains at 15 million/block, the cost of suppressing odd-numbered block transactions is only $77,000/hour at the current ETH price.

In this paper, we showed six different ways to achieve coordination in an adversarial situation where some participants may defect, but coordination can make progress. We also showed how the Ethereum coordination power of smart contracts can be used to promote the common interest of removing EIP-1559. The fact that EIP-1559 benefits inactive users and harms active users makes it unstable because inactive users may not be paying attention to the state of the blockchain when they start cooperating to remove EIP-1559.

Source: medium

By Sergio Demian Lerner

Compiled by: Chen Yiwanfeng