validator.md

Phase 0 -- Honest Validator

This is an accompanying document to Phase 0 -- The Beacon Chain, which describes the expected actions of a "validator" participating in the Ethereum proof-of-stake protocol.

Table of contents

• Introduction
• Prerequisites
• Constants
  • Misc
• Containers
  • Eth1Block
  • AggregateAndProof
  • SignedAggregateAndProof
• Becoming a validator
  • Initialization
    • BLS public key
    • Withdrawal credentials
      • BLS_WITHDRAWAL_PREFIX
      • ETH1_ADDRESS_WITHDRAWAL_PREFIX
  • Submit deposit
  • Process deposit
  • Validator index
  • Activation
• Validator assignments
  • Lookahead
• Beacon chain responsibilities
  • Block proposal
    • Preparing for a BeaconBlock
      • Slot
      • Proposer index
      • Parent root
    • Constructing the BeaconBlockBody
      • Randao reveal
      • Eth1 Data
        • get_eth1_data
      • Proposer slashings
      • Attester slashings
      • Attestations
      • Deposits
      • Voluntary exits
    • Packaging into a SignedBeaconBlock
      • State root
      • Signature
  • Attesting
    • Attestation data
      • General
      • LMD GHOST vote
      • FFG vote
    • Construct attestation
      • Data
      • Aggregation bits
      • Aggregate signature
    • Broadcast attestation
  • Attestation aggregation
    • Aggregation selection
    • Construct aggregate
      • Data
      • Aggregation bits
      • Aggregate signature
    • Broadcast aggregate
• How to avoid slashing
  • Proposer slashing
  • Attester slashing
• Protection best practices

Introduction

This document represents the expected behavior of an "honest validator" with respect to Phase 0 of the Ethereum proof-of-stake protocol. This document does not distinguish between a "node" (i.e. the functionality of following and reading the beacon chain) and a "validator client" (i.e. the functionality of actively participating in consensus). The separation of concerns between these (potentially) two pieces of software is left as a design decision that is out of scope.

A validator is an entity that participates in the consensus of the Ethereum proof-of-stake protocol. This is an optional role for users in which they can post ETH as collateral and verify and attest to the validity of blocks to seek financial returns in exchange for building and securing the protocol. This is similar to proof-of-work networks in which miners provide collateral in the form of hardware/hash-power to seek returns in exchange for building and securing the protocol.

Prerequisites

All terminology, constants, functions, and protocol mechanics defined in the Phase 0 -- The Beacon Chain and Phase 0 -- Deposit Contract doc are requisite for this document and used throughout. Please see the Phase 0 doc before continuing and use as a reference throughout.

Constants

Misc

Name	Value	Unit	Duration
TARGET_AGGREGATORS_PER_COMMITTEE	2**4 (= 16)	validators

Containers

Eth1Block

class Eth1Block(Container):
    timestamp: uint64
    deposit_root: Root
    deposit_count: uint64
    # All other eth1 block fields

AggregateAndProof

class AggregateAndProof(Container):
    aggregator_index: ValidatorIndex
    aggregate: Attestation
    selection_proof: BLSSignature

SignedAggregateAndProof

class SignedAggregateAndProof(Container):
    message: AggregateAndProof
    signature: BLSSignature

Becoming a validator

Initialization

A validator must initialize many parameters locally before submitting a deposit and joining the validator registry.

BLS public key

Validator public keys are G1 points on the BLS12-381 curve. A private key, privkey, must be securely generated along with the resultant pubkey. This privkey must be "hot", that is, constantly available to sign data throughout the lifetime of the validator.

Withdrawal credentials

The withdrawal_credentials field constrains validator withdrawals. The first byte of this 32-byte field is a withdrawal prefix which defines the semantics of the remaining 31 bytes.

The following withdrawal prefixes are currently supported.

BLS_WITHDRAWAL_PREFIX

Withdrawal credentials with the BLS withdrawal prefix allow a BLS key pair (bls_withdrawal_privkey, bls_withdrawal_pubkey) to trigger withdrawals. The withdrawal_credentials field must be such that:

• withdrawal_credentials[:1] == BLS_WITHDRAWAL_PREFIX
• withdrawal_credentials[1:] == hash(bls_withdrawal_pubkey)[1:]

Note: The bls_withdrawal_privkey is not required for validating and can be kept in cold storage.

ETH1_ADDRESS_WITHDRAWAL_PREFIX

Withdrawal credentials with the Eth1 address withdrawal prefix specify a 20-byte Eth1 address eth1_withdrawal_address as the recipient for all withdrawals. The eth1_withdrawal_address can be the address of either an externally owned account or of a contract.

The withdrawal_credentials field must be such that:

• withdrawal_credentials[:1] == ETH1_ADDRESS_WITHDRAWAL_PREFIX
• withdrawal_credentials[1:12] == b'\x00' * 11
• withdrawal_credentials[12:] == eth1_withdrawal_address

After the merge of the current Ethereum execution layer into the Beacon Chain, withdrawals to eth1_withdrawal_address will simply be increases to the account's ETH balance that do NOT trigger any EVM execution.

Submit deposit

In Phase 0, all incoming validator deposits originate from the Ethereum proof-of-work chain defined by DEPOSIT_CHAIN_ID and DEPOSIT_NETWORK_ID. Deposits are made to the deposit contract located at DEPOSIT_CONTRACT_ADDRESS.

To submit a deposit:

• Pack the validator's initialization parameters into deposit_data, a DepositData SSZ object.
• Let amount be the amount in Gwei to be deposited by the validator where amount >= MIN_DEPOSIT_AMOUNT.
• Set deposit_data.pubkey to validator's pubkey.
• Set deposit_data.withdrawal_credentials to withdrawal_credentials.
• Set deposit_data.amount to amount.
• Let deposit_message be a DepositMessage with all the DepositData contents except the signature.
• Let signature be the result of bls.Sign of the compute_signing_root(deposit_message, domain) with domain=compute_domain(DOMAIN_DEPOSIT). (Warning: Deposits must be signed with GENESIS_FORK_VERSION, calling compute_domain without a second argument defaults to the correct version).
• Let deposit_data_root be hash_tree_root(deposit_data).
• Send a transaction on the Ethereum proof-of-work chain to DEPOSIT_CONTRACT_ADDRESS executing def deposit(pubkey: bytes[48], withdrawal_credentials: bytes[32], signature: bytes[96], deposit_data_root: bytes32) along with a deposit of amount Gwei.

Note: Deposits made for the same pubkey are treated as for the same validator. A singular Validator will be added to state.validators with each additional deposit amount added to the validator's balance. A validator can only be activated when total deposits for the validator pubkey meet or exceed MAX_EFFECTIVE_BALANCE.

Process deposit

Deposits cannot be processed into the beacon chain until the proof-of-work block in which they were deposited or any of its descendants is added to the beacon chain state.eth1_data. This takes a minimum of ETH1_FOLLOW_DISTANCE Eth1 blocks (~8 hours) plus EPOCHS_PER_ETH1_VOTING_PERIOD epochs (~6.8 hours). Once the requisite proof-of-work block data is added, the deposit will normally be added to a beacon chain block and processed into the state.validators within an epoch or two. The validator is then in a queue to be activated.

Validator index

Once a validator has been processed and added to the beacon state's validators, the validator's validator_index is defined by the index into the registry at which the ValidatorRecord contains the pubkey specified in the validator's deposit. A validator's validator_index is guaranteed to not change from the time of initial deposit until the validator exits and fully withdraws. This validator_index is used throughout the specification to dictate validator roles and responsibilities at any point and should be stored locally.

Activation

In normal operation, the validator is quickly activated, at which point the validator is added to the shuffling and begins validation after an additional MAX_SEED_LOOKAHEAD epochs (25.6 minutes).

The function is_active_validator can be used to check if a validator is active during a given epoch. Usage is as follows:

def check_if_validator_active(state: BeaconState, validator_index: ValidatorIndex) -> bool:
    validator = state.validators[validator_index]
    return is_active_validator(validator, get_current_epoch(state))

Once a validator is activated, the validator is assigned responsibilities until exited.

Note: There is a maximum validator churn per finalized epoch, so the delay until activation is variable depending upon finality, total active validator balance, and the number of validators in the queue to be activated.

Validator assignments

A validator can get committee assignments for a given epoch using the following helper via get_committee_assignment(state, epoch, validator_index) where epoch <= next_epoch.

def get_committee_assignment(state: BeaconState,
                             epoch: Epoch,
                             validator_index: ValidatorIndex
                             ) -> Optional[Tuple[Sequence[ValidatorIndex], CommitteeIndex, Slot]]:
    """
    Return the committee assignment in the ``epoch`` for ``validator_index``.
    ``assignment`` returned is a tuple of the following form:
        * ``assignment[0]`` is the list of validators in the committee
        * ``assignment[1]`` is the index to which the committee is assigned
        * ``assignment[2]`` is the slot at which the committee is assigned
    Return None if no assignment.
    """
    next_epoch = Epoch(get_current_epoch(state) + 1)
    assert epoch <= next_epoch

    start_slot = compute_start_slot_at_epoch(epoch)
    committee_count_per_slot = get_committee_count_per_slot(state, epoch)
    for slot in range(start_slot, start_slot + SLOTS_PER_EPOCH):
        for index in range(committee_count_per_slot):
            committee = get_beacon_committee(state, Slot(slot), CommitteeIndex(index))
            if validator_index in committee:
                return committee, CommitteeIndex(index), Slot(slot)
    return None

A validator can use the following function to see if they are supposed to propose during a slot. This function can only be run with a state of the slot in question. Proposer selection is only stable within the context of the current epoch.

def is_proposer(state: BeaconState, validator_index: ValidatorIndex) -> bool:
    return get_beacon_proposer_index(state) == validator_index

Note: To see if a validator is assigned to propose during the slot, the beacon state must be in the epoch in question. At the epoch boundaries, the validator must run an epoch transition into the epoch to successfully check the proposal assignment of the first slot.

Note: BeaconBlock proposal is distinct from beacon committee assignment, and in a given epoch each responsibility might occur at a different slot.

Lookahead

The beacon chain shufflings are designed to provide a minimum of 1 epoch lookahead on the validator's upcoming committee assignments for attesting dictated by the shuffling and slot. Note that this lookahead does not apply to proposing, which must be checked during the epoch in question.

get_committee_assignment should be called at the start of each epoch to get the assignment for the next epoch (current_epoch + 1). A validator should plan for future assignments by noting their assigned attestation slot and joining the committee index attestation subnet related to their committee assignment.

Specifically a validator should:

• Call _, committee_index, _ = get_committee_assignment(state, next_epoch, validator_index) when checking for next epoch assignments.
• Calculate the committees per slot for the next epoch: committees_per_slot = get_committee_count_per_slot(state, next_epoch)
• Calculate the subnet index: subnet_id = compute_subnet_for_attestation(committees_per_slot, slot, committee_index)
• Find peers of the pubsub topic beacon_attestation_{subnet_id}.
  • If an insufficient number of current peers are subscribed to the topic, the validator must discover new peers on this topic. Via the discovery protocol, find peers with an ENR containing the attnets entry such that ENR["attnets"][subnet_id] == True. Then validate that the peers are still persisted on the desired topic by requesting GetMetaData and checking the resulting attnets field.
  • If the validator is assigned to be an aggregator for the slot (see is_aggregator()), then subscribe to the topic.

Note: If the validator is not assigned to be an aggregator, the validator only needs sufficient number of peers on the topic to be able to publish messages. The validator does not need to subscribe and listen to all messages on the topic.

Beacon chain responsibilities

A validator has two primary responsibilities to the beacon chain: proposing blocks and creating attestations. Proposals happen infrequently, whereas attestations should be created once per epoch.

Block proposal

A validator is expected to propose a SignedBeaconBlock at the beginning of any slot during which is_proposer(state, validator_index) returns True.

To propose, the validator selects the BeaconBlock, parent which:

1. In their view of fork choice is the head of the chain at the start of slot, after running on_tick and applying any queued attestations from slot - 1.
2. Is from a slot strictly less than the slot of the block about to be proposed, i.e. parent.slot < slot.

The validator creates, signs, and broadcasts a block that is a child of parent that satisfies a valid beacon chain state transition.

There is one proposer per slot, so if there are N active validators any individual validator will on average be assigned to propose once per N slots (e.g. at 312,500 validators = 10 million ETH, that's once per ~6 weeks).

Note: In this section, state is the state of the slot for the block proposal without the block yet applied. That is, state is the previous_state processed through any empty slots up to the assigned slot using process_slots(previous_state, slot).

Preparing for a BeaconBlock

To construct a BeaconBlockBody, a block (BeaconBlock) is defined with the necessary context for a block proposal:

Slot

Set block.slot = slot where slot is the current slot at which the validator has been selected to propose. The parent selected must satisfy that parent.slot < block.slot.

Note: There might be "skipped" slots between the parent and block. These skipped slots are processed in the state transition function without per-block processing.

Proposer index

Set block.proposer_index = validator_index where validator_index is the validator chosen to propose at this slot. The private key mapping to state.validators[validator_index].pubkey is used to sign the block.

Parent root

Set block.parent_root = hash_tree_root(parent).

Constructing the BeaconBlockBody

Randao reveal

Set block.body.randao_reveal = epoch_signature where epoch_signature is obtained from:

def get_epoch_signature(state: BeaconState, block: BeaconBlock, privkey: int) -> BLSSignature:
    domain = get_domain(state, DOMAIN_RANDAO, compute_epoch_at_slot(block.slot))
    signing_root = compute_signing_root(compute_epoch_at_slot(block.slot), domain)
    return bls.Sign(privkey, signing_root)

Eth1 Data

The block.body.eth1_data field is for block proposers to vote on recent Eth1 data. This recent data contains an Eth1 block hash as well as the associated deposit root (as calculated by the get_deposit_root() method of the deposit contract) and deposit count after execution of the corresponding Eth1 block. If over half of the block proposers in the current Eth1 voting period vote for the same eth1_data then state.eth1_data updates immediately allowing new deposits to be processed. Each deposit in block.body.deposits must verify against state.eth1_data.eth1_deposit_root.

get_eth1_data

Let Eth1Block be an abstract object representing Eth1 blocks with the timestamp and deposit contract data available.

Let get_eth1_data(block: Eth1Block) -> Eth1Data be the function that returns the Eth1 data for a given Eth1 block.

An honest block proposer sets block.body.eth1_data = get_eth1_vote(state, eth1_chain) where:

def compute_time_at_slot(state: BeaconState, slot: Slot) -> uint64:
    return uint64(state.genesis_time + slot * SECONDS_PER_SLOT)

def voting_period_start_time(state: BeaconState) -> uint64:
    eth1_voting_period_start_slot = Slot(state.slot - state.slot % (EPOCHS_PER_ETH1_VOTING_PERIOD * SLOTS_PER_EPOCH))
    return compute_time_at_slot(state, eth1_voting_period_start_slot)

def is_candidate_block(block: Eth1Block, period_start: uint64) -> bool:
    return (
        block.timestamp + SECONDS_PER_ETH1_BLOCK * ETH1_FOLLOW_DISTANCE <= period_start
        and block.timestamp + SECONDS_PER_ETH1_BLOCK * ETH1_FOLLOW_DISTANCE * 2 >= period_start
    )

def get_eth1_vote(state: BeaconState, eth1_chain: Sequence[Eth1Block]) -> Eth1Data:
    period_start = voting_period_start_time(state)
    # `eth1_chain` abstractly represents all blocks in the eth1 chain sorted by ascending block height
    votes_to_consider = [
        get_eth1_data(block) for block in eth1_chain
        if (
            is_candidate_block(block, period_start)
            # Ensure cannot move back to earlier deposit contract states
            and get_eth1_data(block).deposit_count >= state.eth1_data.deposit_count
        )
    ]

    # Valid votes already cast during this period
    valid_votes = [vote for vote in state.eth1_data_votes if vote in votes_to_consider]

    # Default vote on latest eth1 block data in the period range unless eth1 chain is not live
    # Non-substantive casting for linter
    state_eth1_data: Eth1Data = state.eth1_data
    default_vote = votes_to_consider[len(votes_to_consider) - 1] if any(votes_to_consider) else state_eth1_data

    return max(
        valid_votes,
        key=lambda v: (valid_votes.count(v), -valid_votes.index(v)),  # Tiebreak by smallest distance
        default=default_vote
    )

Proposer slashings

Up to MAX_PROPOSER_SLASHINGS, ProposerSlashing objects can be included in the block. The proposer slashings must satisfy the verification conditions found in proposer slashings processing. The validator receives a small "whistleblower" reward for each proposer slashing found and included.

Attester slashings

Up to MAX_ATTESTER_SLASHINGS, AttesterSlashing objects can be included in the block. The attester slashings must satisfy the verification conditions found in attester slashings processing. The validator receives a small "whistleblower" reward for each attester slashing found and included.

Attestations

Up to MAX_ATTESTATIONS, aggregate attestations can be included in the block. The attestations added must satisfy the verification conditions found in attestation processing. To maximize profit, the validator should attempt to gather aggregate attestations that include singular attestations from the largest number of validators whose signatures from the same epoch have not previously been added on chain.

Deposits

If there are any unprocessed deposits for the existing state.eth1_data (i.e. state.eth1_data.deposit_count > state.eth1_deposit_index), then pending deposits must be added to the block. The expected number of deposits is exactly min(MAX_DEPOSITS, eth1_data.deposit_count - state.eth1_deposit_index). These deposits are constructed from the Deposit logs from the deposit contract and must be processed in sequential order. The deposits included in the block must satisfy the verification conditions found in deposits processing.

The proof for each deposit must be constructed against the deposit root contained in state.eth1_data rather than the deposit root at the time the deposit was initially logged from the proof-of-work chain. This entails storing a full deposit merkle tree locally and computing updated proofs against the eth1_data.deposit_root as needed. See minimal_merkle.py for a sample implementation.

Voluntary exits

Up to MAX_VOLUNTARY_EXITS, VoluntaryExit objects can be included in the block. The exits must satisfy the verification conditions found in exits processing.

Note: If a slashing for a validator is included in the same block as a voluntary exit, the voluntary exit will fail and cause the block to be invalid due to the slashing being processed first. Implementers must take heed of this operation interaction when packing blocks.

Packaging into a SignedBeaconBlock

State root

Set block.state_root = hash_tree_root(state) of the resulting state of the parent -> block state transition.

Note: To calculate state_root, the validator should first run the state transition function on an unsigned block containing a stub for the state_root. It is useful to be able to run a state transition function (working on a copy of the state) that does not validate signatures or state root for this purpose:

def compute_new_state_root(state: BeaconState, block: BeaconBlock) -> Root:
    temp_state: BeaconState = state.copy()
    signed_block = SignedBeaconBlock(message=block)
    state_transition(temp_state, signed_block, validate_result=False)
    return hash_tree_root(temp_state)

Signature

signed_block = SignedBeaconBlock(message=block, signature=block_signature), where block_signature is obtained from:

def get_block_signature(state: BeaconState, block: BeaconBlock, privkey: int) -> BLSSignature:
    domain = get_domain(state, DOMAIN_BEACON_PROPOSER, compute_epoch_at_slot(block.slot))
    signing_root = compute_signing_root(block, domain)
    return bls.Sign(privkey, signing_root)

Attesting

A validator is expected to create, sign, and broadcast an attestation during each epoch. The committee, assigned index, and assigned slot for which the validator performs this role during an epoch are defined by get_committee_assignment(state, epoch, validator_index).

A validator should create and broadcast the attestation to the associated attestation subnet when either (a) the validator has received a valid block from the expected block proposer for the assigned slot or (b) 1 / INTERVALS_PER_SLOT of the slot has transpired (SECONDS_PER_SLOT / INTERVALS_PER_SLOT seconds after the start of slot) -- whichever comes first.

Note: Although attestations during GENESIS_EPOCH do not count toward FFG finality, these initial attestations do give weight to the fork choice, are rewarded, and should be made.

Attestation data

First, the validator should construct attestation_data, an AttestationData object based upon the state at the assigned slot.

• Let head_block be the result of running the fork choice during the assigned slot.
• Let head_state be the state of head_block processed through any empty slots up to the assigned slot using process_slots(state, slot).

General

• Set attestation_data.slot = slot where slot is the assigned slot.
• Set attestation_data.index = index where index is the index associated with the validator's committee.

LMD GHOST vote

Set attestation_data.beacon_block_root = hash_tree_root(head_block).

FFG vote

• Set attestation_data.source = head_state.current_justified_checkpoint.
• Set attestation_data.target = Checkpoint(epoch=get_current_epoch(head_state), root=epoch_boundary_block_root) where epoch_boundary_block_root is the root of block at the most recent epoch boundary.

Note: epoch_boundary_block_root can be looked up in the state using:

• Let start_slot = compute_start_slot_at_epoch(get_current_epoch(head_state)).
• Let epoch_boundary_block_root = hash_tree_root(head_block) if start_slot == head_state.slot else get_block_root(state, get_current_epoch(head_state)).

Construct attestation

Next, the validator creates attestation, an Attestation object.

Data

Set attestation.data = attestation_data where attestation_data is the AttestationData object defined in the previous section, attestation data.

Aggregation bits

• Let attestation.aggregation_bits be a Bitlist[MAX_VALIDATORS_PER_COMMITTEE] of length len(committee), where the bit of the index of the validator in the committee is set to 0b1.

Note: Calling get_attesting_indices(state, attestation.data, attestation.aggregation_bits) should return a list of length equal to 1, containing validator_index.

Aggregate signature

Set attestation.signature = attestation_signature where attestation_signature is obtained from:

def get_attestation_signature(state: BeaconState, attestation_data: AttestationData, privkey: int) -> BLSSignature:
    domain = get_domain(state, DOMAIN_BEACON_ATTESTER, attestation_data.target.epoch)
    signing_root = compute_signing_root(attestation_data, domain)
    return bls.Sign(privkey, signing_root)

Broadcast attestation

Finally, the validator broadcasts attestation to the associated attestation subnet, the beacon_attestation_{subnet_id} pubsub topic.

The subnet_id for the attestation is calculated with:

• Let committees_per_slot = get_committee_count_per_slot(state, attestation.data.target.epoch).
• Let subnet_id = compute_subnet_for_attestation(committees_per_slot, attestation.data.slot, attestation.data.index).

def compute_subnet_for_attestation(committees_per_slot: uint64,
                                   slot: Slot,
                                   committee_index: CommitteeIndex) -> SubnetID:
    """
    Compute the correct subnet for an attestation for Phase 0.
    Note, this mimics expected future behavior where attestations will be mapped to their shard subnet.
    """
    slots_since_epoch_start = uint64(slot % SLOTS_PER_EPOCH)
    committees_since_epoch_start = committees_per_slot * slots_since_epoch_start

    return SubnetID((committees_since_epoch_start + committee_index) % ATTESTATION_SUBNET_COUNT)

Attestation aggregation

Some validators are selected to locally aggregate attestations with a similar attestation_data to their constructed attestation for the assigned slot.

Aggregation selection

A validator is selected to aggregate based upon the return value of is_aggregator().

def get_slot_signature(state: BeaconState, slot: Slot, privkey: int) -> BLSSignature:
    domain = get_domain(state, DOMAIN_SELECTION_PROOF, compute_epoch_at_slot(slot))
    signing_root = compute_signing_root(slot, domain)
    return bls.Sign(privkey, signing_root)

def is_aggregator(state: BeaconState, slot: Slot, index: CommitteeIndex, slot_signature: BLSSignature) -> bool:
    committee = get_beacon_committee(state, slot, index)
    modulo = max(1, len(committee) // TARGET_AGGREGATORS_PER_COMMITTEE)
    return bytes_to_uint64(hash(slot_signature)[0:8]) % modulo == 0

Construct aggregate

If the validator is selected to aggregate (is_aggregator()), they construct an aggregate attestation via the following.

Collect attestations seen via gossip during the slot that have an equivalent attestation_data to that constructed by the validator. If len(attestations) > 0, create an aggregate_attestation: Attestation with the following fields.

Data

Set aggregate_attestation.data = attestation_data where attestation_data is the AttestationData object that is the same for each individual attestation being aggregated.

Aggregation bits

Let aggregate_attestation.aggregation_bits be a Bitlist[MAX_VALIDATORS_PER_COMMITTEE] of length len(committee), where each bit set from each individual attestation is set to 0b1.

Aggregate signature

Set aggregate_attestation.signature = aggregate_signature where aggregate_signature is obtained from:

def get_aggregate_signature(attestations: Sequence[Attestation]) -> BLSSignature:
    signatures = [attestation.signature for attestation in attestations]
    return bls.Aggregate(signatures)

Broadcast aggregate

If the validator is selected to aggregate (is_aggregator), then they broadcast their best aggregate as a SignedAggregateAndProof to the global aggregate channel (beacon_aggregate_and_proof) 2 / INTERVALS_PER_SLOT of the way through the slot-that is, SECONDS_PER_SLOT * 2 / INTERVALS_PER_SLOT seconds after the start of slot.

Selection proofs are provided in AggregateAndProof to prove to the gossip channel that the validator has been selected as an aggregator.

AggregateAndProof messages are signed by the aggregator and broadcast inside of SignedAggregateAndProof objects to prevent a class of DoS attacks and message forgeries.

First, aggregate_and_proof = get_aggregate_and_proof(state, validator_index, aggregate_attestation, privkey) is constructed.

def get_aggregate_and_proof(state: BeaconState,
                            aggregator_index: ValidatorIndex,
                            aggregate: Attestation,
                            privkey: int) -> AggregateAndProof:
    return AggregateAndProof(
        aggregator_index=aggregator_index,
        aggregate=aggregate,
        selection_proof=get_slot_signature(state, aggregate.data.slot, privkey),
    )

Then signed_aggregate_and_proof = SignedAggregateAndProof(message=aggregate_and_proof, signature=signature) is constructed and broadcast. Where signature is obtained from:

def get_aggregate_and_proof_signature(state: BeaconState,
                                      aggregate_and_proof: AggregateAndProof,
                                      privkey: int) -> BLSSignature:
    aggregate = aggregate_and_proof.aggregate
    domain = get_domain(state, DOMAIN_AGGREGATE_AND_PROOF, compute_epoch_at_slot(aggregate.data.slot))
    signing_root = compute_signing_root(aggregate_and_proof, domain)
    return bls.Sign(privkey, signing_root)

How to avoid slashing

"Slashing" is the burning of some amount of validator funds and immediate ejection from the active validator set. In Phase 0, there are two ways in which funds can be slashed: proposer slashing and attester slashing. Although being slashed has serious repercussions, it is simple enough to avoid being slashed all together by remaining consistent with respect to the messages a validator has previously signed.

Note: Signed data must be within a sequential Fork context to conflict. Messages cannot be slashed across diverging forks. If the previous fork version is 1 and the chain splits into fork 2 and 102, messages from 1 can be slashable against messages in forks 1, 2, and 102. Messages in 2 cannot be slashable against messages in 102, and vice versa.

Proposer slashing

To avoid "proposer slashings", a validator must not sign two conflicting BeaconBlock where conflicting is defined as two distinct blocks within the same slot.

In Phase 0, as long as the validator does not sign two different beacon blocks for the same slot, the validator is safe against proposer slashings.

Specifically, when signing a BeaconBlock, a validator should perform the following steps in the following order:

1. Save a record to hard disk that a beacon block has been signed for the slot=block.slot.
2. Generate and broadcast the block.

If the software crashes at some point within this routine, then when the validator comes back online, the hard disk has the record of the potentially signed/broadcast block and can effectively avoid slashing.

Attester slashing

To avoid "attester slashings", a validator must not sign two conflicting AttestationData objects, i.e. two attestations that satisfy is_slashable_attestation_data.

Specifically, when signing an Attestation, a validator should perform the following steps in the following order:

1. Save a record to hard disk that an attestation has been signed for source (i.e. attestation_data.source.epoch) and target (i.e. attestation_data.target.epoch).
2. Generate and broadcast attestation.

If the software crashes at some point within this routine, then when the validator comes back online, the hard disk has the record of the potentially signed/broadcast attestation and can effectively avoid slashing.

Protection best practices

A validator client should be considered standalone and should consider the beacon node as untrusted. This means that the validator client should protect:

1. Private keys -- private keys should be protected from being exported accidentally or by an attacker.
2. Slashing -- before a validator client signs a message it should validate the data, check it against a local slashing database (do not sign a slashable attestation or block) and update its internal slashing database with the newly signed object.
3. Recovered validator -- Recovering a validator from a private key will result in an empty local slashing db. Best practice is to import (from a trusted source) that validator's attestation history. See EIP 3076 for a standard slashing interchange format.
4. Far future signing requests -- A validator client can be requested to sign a far into the future attestation, resulting in a valid non-slashable request. If the validator client signs this message, it will result in it blocking itself from attesting any other attestation until the beacon-chain reaches that far into the future epoch. This will result in an inactivity penalty and potential ejection due to low balance. A validator client should prevent itself from signing such requests by: a) keeping a local time clock if possible and following best practices to stop time server attacks and b) refusing to sign, by default, any message that has a large (>6h) gap from the current slashing protection database indicated a time "jump" or a long offline event. The administrator can manually override this protection to restart the validator after a genuine long offline event.