Skip to content

Latest commit

 

History

History
609 lines (501 loc) · 22.5 KB

eip-7251.md

File metadata and controls

609 lines (501 loc) · 22.5 KB
Raw
eip title description author discussions-to status type category created requires
7251
Increase the MAX_EFFECTIVE_BALANCE
Allow validators to have larger effective balances, while maintaining the 32 ETH lower bound.
mike (@michaelneuder), Francesco (@fradamt), dapplion (@dapplion), Mikhail (@mkalinin), Aditya (@adiasg), Justin (@justindrake), lightclient (@lightclient)
Review
Standards Track
Core
2023-06-28
7002, 7685

Abstract

Increases the constant MAX_EFFECTIVE_BALANCE, while keeping the minimum staking balance 32 ETH. This permits large node operators to consolidate into fewer validators while also allowing solo-stakers to earn compounding rewards and stake in more flexible increments.

Motivation

As of October 3, 2023, there are currently over 830,000 validators participating in the consensus layer. The size of this set continues to grow due, in part, to the MAX_EFFECTIVE_BALANCE, which limits the stake of a single validator to 32 ETH. This leads to large amounts of "redundant validators", which are controlled by a single entity, possibly running on the same beacon node, but with distinct BLS signing keys. The limit on the MAX_EFFECTIVE_BALANCE is technical debt from the original sharding design, in which subcommittees (not the attesting committee but the committee calculated in is_aggregator) needed to be majority honest. As a result, keeping the weights of subcommittee members approximately equal reduced the risk of a single large validator containing too much influence. Under the current design, these subcommittees are only used for attestation aggregation, and thus only have a 1/N honesty assumption.

With the security model of the protocol no longer dependent on a low value for MAX_EFFECTIVE_BALANCE, we propose raising this value while keeping the minimum validator threshold of 32 ETH. This increase aims to reduce the validator set size, thereby reducing the number of P2P messages over the network, the number of BLS signatures that need to be aggregated each epoch, and the BeaconState memory footprint. This change adds value for both small and large validators. Large validators can consolidate to run fewer validators and thus fewer beacon nodes. Small validators now benefit from compounding rewards and the ability to stake in more flexible increments (e.g., the ability to stake 40 ETH instead of needing to accumulate 64 ETH to run two validators today).

Specification

Constants

Execution layer

Name Value Comment
CONSOLIDATION_REQUEST_TYPE 0x02 The EIP-7685 type prefix for consolidation request
CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS 0x Where to call and store relevant details about consolidation requests mechanism
SYSTEM_ADDRESS 0xfffffffffffffffffffffffffffffffffffffffe Address used to invoke system operation on contract
EXCESS_CONSOLIDATION_REQUESTS_STORAGE_SLOT 0
CONSOLIDATION_REQUEST_COUNT_STORAGE_SLOT 1
CONSOLIDATION_REQUEST_QUEUE_HEAD_STORAGE_SLOT 2 Pointer to head of the consolidation request message queue
CONSOLIDATION_REQUEST_QUEUE_TAIL_STORAGE_SLOT 3 Pointer to the tail of the consolidation request message queue
CONSOLIDATION_REQUEST_QUEUE_STORAGE_OFFSET 4 The start memory slot of the in-state consolidation request message queue
MAX_CONSOLIDATION_REQUESTS_PER_BLOCK 1 Maximum number of consolidation requests that can be dequeued into a block
TARGET_CONSOLIDATION_REQUESTS_PER_BLOCK 1
MIN_CONSOLIDATION_REQUEST_FEE 1
CONSOLIDATION_REQUEST_FEE_UPDATE_FRACTION 17
EXCESS_INHIBITOR 1181 Excess value used to compute the fee before the first system call
FORK_TIMESTAMP TBD Mainnet

Consensus layer

Name Value
COMPOUNDING_WITHDRAWAL_PREFIX Bytes1('0x02')
MIN_ACTIVATION_BALANCE Gwei(2**5 * 10**9) (32 ETH)
MAX_EFFECTIVE_BALANCE Gwei(2**11 * 10**9) (2048 ETH)

Execution layer

Consolidation request

The new consolidation request is an EIP-7685 request with type 0x02 consisting of the following fields:

  1. source_address: Bytes20
  2. source_pubkey: Bytes48
  3. target_pubkey: Bytes48

The EIP-7685 encoding of a withdrawal request MUST be computed as the follows:

encoded = CONSOLIDATION_REQUEST_TYPE ++ rlp([source_address, source_pubkey, target_pubkey])

Consolidation request contract

The contract has three different code paths, which can be summarized at a high level as follows:

  1. Add consolidation request - requires a 96 byte input, concatenated public keys of the source and the target validators.
  2. Excess consolidation requests getter - if the input length is zero, return the current excess consolidation requests count.
  3. System process - if called by system address, pop off the consolidation requests for the current block from the queue.
Add Consolidation Request

If call data input to the contract is exactly 96 bytes, perform the following:

  1. Ensure enough ETH was sent to cover the current consolidation request fee (check_fee())
  2. Increase consolidation request count by 1 for the current block (increment_count())
  3. Insert a consolidation request into the queue for the source address and pubkeys of the source and the target (insert_withdrawal_request_into_queue())

Specifically, the functionality is defined in pseudocode as the function add_consolidation_request():

def add_consolidation_request(Bytes48: source_pubkey, Bytes48: target_pubkey):
    """
    Add consolidaiton request adds new request to the consolidation request queue, so long as a sufficient fee is provided.
    """

    # Verify sufficient fee was provided.
    fee = get_fee()
    require(msg.value >= fee, 'Insufficient value for fee')

    # Increment consolidaiton request count.
    count = sload(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, CONSOLIDATION_REQUEST_COUNT_STORAGE_SLOT)
    sstore(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, CONSOLIDATION_REQUEST_COUNT_STORAGE_SLOT, count + 1)

    # Insert into queue.
    queue_tail_index = sload(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, CONSOLIDATION_REQUEST_QUEUE_TAIL_STORAGE_SLOT)
    queue_storage_slot = CONSOLIDATION_REQUEST_QUEUE_STORAGE_OFFSET + queue_tail_index * 4
    sstore(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, queue_storage_slot, msg.sender)
    sstore(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, queue_storage_slot + 1, source_pubkey[0:32])
    sstore(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, queue_storage_slot + 2, source_pubkey[32:48] ++ target_pubkey[0:16])
    sstore(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, queue_storage_slot + 3, target_pubkey[16:48])
    sstore(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, CONSOLIDATION_REQUEST_QUEUE_TAIL_STORAGE_SLOT, queue_tail_index + 1)
Fee calculation

The following pseudocode can compute the cost an individual consolidation request, given a certain number of excess consolidation requests.

def get_fee() -> int:
    excess = sload(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, EXCESS_CONSOLIDATION_REQUESTS_STORAGE_SLOT)
    return fake_exponential(
        MIN_CONSOLIDATION_REQUEST_FEE,
        excess,
        CONSOLIDATION_REQUEST_FEE_UPDATE_FRACTION
    )

def fake_exponential(factor: int, numerator: int, denominator: int) -> int:
    i = 1
    output = 0
    numerator_accum = factor * denominator
    while numerator_accum > 0:
        output += numerator_accum
        numerator_accum = (numerator_accum * numerator) // (denominator * i)
        i += 1
    return output // denominator
Excess Consolidation Requests Getter
def get_excess_consolidation_requests():
    count = sload(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, EXCESS_CONSOLIDATION_REQUESTS_STORAGE_SLOT)
    return count
System Call

If the contract is called as SYSTEM_ADDRESS with an empty input data, perform the following:

  • The contract's queue is updated based on consolidation requests dequeued and the consolidation requests queue head/tail are reset if the queue has been cleared (dequeue_consolidation_requests())
  • The contract's excess consolidation requests are updated based on usage in the current block (update_excess_consolidation_requests())
  • The contract's consolidation requests count is reset to 0 (reset_consolidation_requests_count())

Specifically, the functionality is defined in pseudocode as the function process_consolidation_requests():

###################
# Public function #
###################

def process_consolidation_requests():
    reqs = dequeue_consolidation_requests()
    update_excess_consolidation_requests()
    reset_consolidation_requests_count()
    return reqs

###########
# Helpers #
###########

class ConsolidationRequest(object):
    source_address: Bytes20
    source_pubkey: Bytes48
    target_pubkey: Bytes48

def dequeue_consolidation_requests():
    queue_head_index = sload(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, CONSOLIDATION_REQUEST_QUEUE_HEAD_STORAGE_SLOT)
    queue_tail_index = sload(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, CONSOLIDATION_REQUEST_QUEUE_TAIL_STORAGE_SLOT)
    num_in_queue = queue_tail_index - queue_head_index
    num_dequeued = min(num_in_queue, MAX_CONSOLIDATION_REQUESTS_PER_BLOCK)

    reqs = []
    for i in range(num_dequeue):
        queue_storage_slot = CONSOLIDATION_REQUEST_QUEUE_STORAGE_OFFSET + (queue_head_index + i) * 4
        source_address = address(sload(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, queue_storage_slot)[0:20])
        source_pubkey = (
            sload(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, queue_storage_slot + 1)[0:32] + sload(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, queue_storage_slot + 2)[0:16]
        )
        target_pubkey = (
            sload(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, queue_storage_slot + 2)[16:32] + sload(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, queue_storage_slot + 3)[0:32]
        )
        req = ConsolidationRequest(
            source_address=Bytes20(source_address),
            source_pubkey=Bytes48(source_pubkey),
            target_pubkey=Bytes48(target_pubkey)
        )
        reqs.append(req)

    new_queue_head_index = queue_head_index + num_dequeued
    if new_queue_head_index == queue_tail_index:
        # Queue is empty, reset queue pointers
        sstore(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, CONSOLIDATION_REQUEST_QUEUE_HEAD_STORAGE_SLOT, 0)
        sstore(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, CONSOLIDATION_REQUEST_QUEUE_TAIL_STORAGE_SLOT, 0)
    else:
        sstore(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, CONSOLIDATION_REQUEST_QUEUE_HEAD_STORAGE_SLOT, new_queue_head_index)

    return reqs

def update_excess_consolidation_requests():
    previous_excess = sload(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, EXCESS_CONSOLIDATION_REQUESTS_STORAGE_SLOT)
    # Check if excess needs to be reset to 0 for first iteration after activation
    if previous_excess == EXCESS_INHIBITOR:
        previous_excess = 0

    count = sload(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, CONSOLIDATION_REQUEST_COUNT_STORAGE_SLOT)

    new_excess = 0
    if previous_excess + count > TARGET_CONSOLIDATION_REQUESTS_PER_BLOCK:
        new_excess = previous_excess + count - TARGET_CONSOLIDATION_REQUESTS_PER_BLOCK

    sstore(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, EXCESS_CONSOLIDATION_REQUESTS_STORAGE_SLOT, new_excess)

def reset_consolidation_requests_count():
    sstore(CONSOLIDATION_REQUEST_PREDEPLOY_ADDRESS, CONSOLIDATION_REQUEST_COUNT_STORAGE_SLOT, 0)
Bytecode
caller
push20 0xfffffffffffffffffffffffffffffffffffffffe
eq
push1 0x98
jumpi

calldatasize
iszero
iszero
push1 0x28
jumpi

push0
sload
push0
mstore
push1 0x20
push0
return

jumpdest
calldatasize
push1 0x60
eq
iszero
push2 0x0144
jumpi

push1 0x11
push0
sload
push1 0x01
dup3
mul
push1 0x01
swap1
push0

jumpdest
push0
dup3
gt
iszero
push1 0x59
jumpi

dup2
add
swap1
dup4
mul
dup5
dup4
mul
swap1
div
swap2
push1 0x01
add
swap2
swap1
push1 0x3e
jump

jumpdest
swap1
swap4
swap1
div
callvalue
lt
push2 0x0144
jumpi

push1 0x01
sload
push1 0x01
add
push1 0x01
sstore
push1 0x03
sload
dup1
push1 0x04
mul
push1 0x04
add
caller
dup2
sstore
push1 0x01
add
push0
calldataload
dup2
sstore
push1 0x01
add
push1 0x20
calldataload
dup2
sstore
push1 0x01
add
push1 0x40
calldataload
swap1
sstore
push1 0x01
add
push1 0x03
sstore
stop

jumpdest
push1 0x03
sload
push1 0x02
sload
dup1
dup3
sub
dup1
push1 0x01
gt
push1 0xac
jumpi

pop
push1 0x01

jumpdest
push0

jumpdest
dup2
dup2
eq
push1 0xf1
jumpi

dup1
push1 0x74
mul
dup4
dup3
add
push1 0x04
mul
push1 0x04
add
dup1
sload
swap1
push1 0x01
add
dup1
sload
swap1
push1 0x01
add
dup1
sload
swap1
push1 0x01
add
sload
swap3
push1 0x60
shl
dup5
mstore
swap1
dup4
push1 0x14
add
mstore
dup3
push1 0x34
add
mstore
swap1
push1 0x54
add
mstore
push1 0x01
add
push1 0xae
jump

jumpdest
swap2
add
dup1
swap3
eq
push2 0x0103
jumpi

swap1
push1 0x02
sstore
push2 0x010e
jump

jumpdest
swap1
pop
push0
push1 0x02
sstore
push0
push1 0x03
sstore

jumpdest
push0
sload
dup1
push2 0x049d
eq
iszero
push2 0x011d
jumpi

pop
push0

jumpdest
push1 0x01
sload
push1 0x01
dup3
dup3
add
gt
push2 0x0132
jumpi

pop
pop
push0
push2 0x0138
jump

jumpdest
add
push1 0x01
swap1
sub

jumpdest
push0
sstore
push0
push1 0x01
sstore
push1 0x74
mul
push0
return

jumpdest
push0
push0
revert
Deployment

The consolidation requests contract is deployed like any other smart contract. A special synthetic address is generated by working backwards from the desired deployment transaction:

{
  "type": "0x0",
  "nonce": "0x0",
  "to": null,
  "gas": "0x3d090",
  "gasPrice": "0xe8d4a51000",
  "maxPriorityFeePerGas": null,
  "maxFeePerGas": null,
  "value": "0x0",
  "input": "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",
  "v": "0x1b",
  "r": "0x539",
  "s": "0x13370066aa8fbe21ca1511",
  "hash": "0x6de32a89ba4c0592fd2453f8838d6bb69d93a102723de5f4b0f046ddcf8b8fa9"
}
Sender: 0xd6e25886D7B986C394156C31a48e84Ee0BA71f72
Address: 0x00b42dbF2194e931E80326D950320f7d9Dbeac02

Block processing

At the end of processing any execution block where block.timestamp >= FORK_TIMESTAMP (i.e. after processing all transactions and after performing the block body requests validations) clienst software MUST take the following steps:

  1. Call the contract as SYSTEM_ADDRESS and empty input data to trigger the system subroutine execute.
  2. Check that consolidation requests in the EIP-7685 requests list matches the list returned from dequeue_consolidation_requests() function of the smart contract respecting the order of the returned requests. If this condition does not hold, the block MUST be deemed invalid.

Consensus layer

The defining features of this EIP are:

  1. Increasing the MAX_EFFECTIVE_BALANCE, while creating a MIN_ACTIVATION_BALANCE. The core feature of allowing variable size validators.
  2. Allowing for multiple validator indices to be combined through the protocol. A mechanism by which large node operators can combine validators without cycling through the exit and activation queues.
  3. Permitting validators to set custom ceilings for their validator to indicate where the partial withdrawal sweep activates. Allows more flexibility in defining the "ceiling" of a validator's effective balance.
  4. Adding execution layer partial withdrawals (part of EIP-7002). Allowing Execution Layer messages to trigger partial withdrawals in addition to full exits (e.g., a 100 ETH validator can remove up to 68 ETH without exiting the validator).
  5. Removing the initial slashing penalty (still in discussion). This reduces the risk of consolidation for large validators.

The Rationale section contains an explanation for each of these proposed core features. A sketch of the resulting changes to the consensus layer is included below.

  1. Add COMPOUNDING_WITHDRAWAL_PREFIX and MIN_ACTIVATION_BALANCE constants, while updating the value of MAX_EFFECTIVE_BALANCE.
  2. Create the PendingDeposit container, which is used to track incoming deposits in the weight-based rate limiting mechanism.
  3. Update the BeaconState with fields needed for deposit and exit queue weight-based rate limiting.
  4. Modify is_eligible_for_activation_queue to check against MIN_ACTIVATION_BALANCE rather than MAX_EFFECTIVE_BALANCE.
  5. Modify get_validator_churn_limit to depend on the validator weight rather than the validator count.
  6. Create a helper compute_exit_epoch_and_update_churn to calculate the exit epoch based on the current pending withdrawals.
  7. Modify initiate_validator_exit to rate limit the exit queue by balance rather than the number of validators.
  8. Modify initialize_beacon_state_from_eth1 to use MIN_ACTIVATION_BALANCE.
  9. Modify process_registry_updates to activate all eligible validators.
  10. Add a per-epoch helper, process_pending_balance_deposits, to consume some of the pending deposits.
  11. Modify get_validator_from_deposit to initialize the effective balance to zero (it's updated by the pending deposit flow).
  12. Modify apply_deposit to store incoming deposits in state.pending_balance_deposits.
  13. Modify is_aggregator to be weight-based.
  14. Modify compute_weak_subjectivity_period to use the new churn limit function.
  15. Add has_compounding_withdrawal_credential to check for the 0x02 credential.
  16. Modify is_fully_withdrawable_validator to check for compounding credentials.
  17. Add get_validator_excess_balance to calculate the excess balance of validators.
  18. Modify is_partially_withdrawable_validator to check for excess balance.
  19. Modify get_expected_withdrawals to use excess balance.

Rationale

This EIP aims to reduce the total number of validators without changing anything about the economic security of the protocol. It provides a mechanism by which large node operators who control significant amounts of stake can consolidate into fewer validators. We analyze the reasoning behind each of the core features.

  1. Increasing the MAX_EFFECTIVE_BALANCE, while creating a MIN_ACTIVATION_BALANCE.
    • While increasing the MAX_EFFECTIVE_BALANCE to allow larger-stake validators, it is important to keep the lower bound of 32 ETH (by introducing a new constant – MIN_ACTIVATION_BALANCE) to encourage solo-staking.
  2. Allowing for multiple validator indices to be combined through the protocol.
    • For large staking pools that already control thousands of validators, exiting and re-entering would be extremely slow and costly. The adoption of the EIP will be much higher by allowing in-protocol consolidation.
  3. Permitting validators to set custom ceilings for their validator to indicate where the partial withdrawal sweep activates.
    • To get access to rewards, validators might want the flexibility to set custom ceilings for their effective balance. This gives them more optionality and is a clean way to continue supporting the partial-withdrawal sweep (a gasless way to extract rewards).
  4. Adding execution layer partial withdrawals (part of EIP-7002).
    • For validators that choose to raise their effective balance ceiling, allowing for custom partial withdrawals triggered from the execution layer increases the flexibility of the staking configurations. Validators can choose when and how much they withdraw but will have to pay gas for the EL transaction.
  5. Removing the initial slashing penalty (still in discussion).
    • To encourage consolidation, we could modify the slashing penalties. The biggest hit comes from the initial penalty of 1/32 of the validator's effective balance. Since this scales linearly on the effective balance, the higher-stake validators directly incur higher risk. By changing the scaling properties, we could make consolidation more attractive.

Backwards Compatibility

This EIP introduces backward incompatible changes to the block validation rule set on the consensus layer and must be accompanied by a hard fork. These changes do not break anything related to current user activity and experience.

Security Considerations

This change modifies committees and churn, but doesn't significantly impact the security properties.

Security of attestation committees

Given full consolidation as the worst case, the probability of an adversarial takeover of a committee remains low. Even in a high consolidation scenario, the required share of honest validators remains well below the 2/3 supermajority needed for finality.

Aggregator selection

In the original sharding roadmap, subcommittees were required to be secure with extremely high probability. Now with the sole responsibility of attestation aggregation, we only require each committee to have at least one honest aggregator. Currently, aggregators are selected through a VRF lottery, targeting several validator units that can be biased by non-consolidated attackers. This proposal changes the VRF lottery to consider weight, so the probability of having at least one honest aggregator is not worse.

Proposer selection probability

Proposer selection is already weighted by the ratio of their effective balance to MAX_EFFECTIVE_BALANCE. Due to the lower probabilities, this change will slightly increase the time it takes to calculate the next proposer index.

Sync committee selection probability

Sync committee selection is also already weighted by effective balance, so this proposal does not require modifications to the sync protocol. Light clients can still check that a super-majority of participants have signed an update irrespective of their weights since we maintain a weight-based selection probability.

Churn invariants

This proposal maintains the activation and exit churn invariants limiting active weight instead of validator count. Balance top-ups are now handled explicitly, being subject to the same activation queue as full deposits.

Copyright

Copyright and related rights waived via CC0.