Merge pull request #1876 from ethereum/val-subnet-fix
Fix attestation subnet lookahead
This commit is contained in:
commit
e6d8a8641d
|
@ -275,7 +275,7 @@ Additional global topics are used to propagate lower frequency validator message
|
|||
Attestation subnets are used to propagate unaggregated attestations to subsections of the network. Their `Name`s are:
|
||||
|
||||
- `beacon_attestation_{subnet_id}` - These topics are used to propagate unaggregated attestations to the subnet `subnet_id` (typically beacon and persistent committees) to be aggregated before being gossiped to `beacon_aggregate_and_proof`. The following validations MUST pass before forwarding the `attestation` on the subnet.
|
||||
- _[REJECT]_ The attestation is for the correct subnet (i.e. `compute_subnet_for_attestation(state, attestation) == subnet_id`).
|
||||
- _[REJECT]_ The attestation is for the correct subnet (i.e. `compute_subnet_for_attestation(state, attestation.data.slot, attestation.data.index) == subnet_id`).
|
||||
- _[IGNORE]_ `attestation.data.slot` is within the last `ATTESTATION_PROPAGATION_SLOT_RANGE` slots (within a `MAXIMUM_GOSSIP_CLOCK_DISPARITY` allowance) -- i.e. `attestation.data.slot + ATTESTATION_PROPAGATION_SLOT_RANGE >= current_slot >= attestation.data.slot` (a client MAY queue future attestations for processing at the appropriate slot).
|
||||
- _[REJECT]_ The attestation is unaggregated -- that is, it has exactly one participating validator (`len(get_attesting_indices(state, attestation.data, attestation.aggregation_bits)) == 1`).
|
||||
- _[IGNORE]_ There has been no other valid attestation seen on an attestation subnet that has an identical `attestation.data.target.epoch` and participating validator index.
|
||||
|
@ -286,7 +286,7 @@ Attestation subnets are used to propagate unaggregated attestations to subsectio
|
|||
|
||||
Attestation broadcasting is grouped into subnets defined by a topic. The number of subnets is defined via `ATTESTATION_SUBNET_COUNT`. The correct subnet for an attestation can be calculated with `compute_subnet_for_attestation`. `beacon_attestation_{subnet_id}` topics, are rotated through throughout the epoch in a similar fashion to rotating through shards in committees in Phase 1.
|
||||
|
||||
Unaggregated attestations are sent to the subnet topic, `beacon_attestation_{compute_subnet_for_attestation(state, attestation)}` as `Attestation`s.
|
||||
Unaggregated attestations are sent to the subnet topic, `beacon_attestation_{compute_subnet_for_attestation(state, attestation.data.slot, attestation.data.index)}` as `Attestation`s.
|
||||
|
||||
Aggregated attestations are sent to the `beacon_aggregate_and_proof` topic as `AggregateAndProof`s.
|
||||
|
||||
|
|
|
@ -199,8 +199,8 @@ The beacon chain shufflings are designed to provide a minimum of 1 epoch lookahe
|
|||
|
||||
Specifically a validator should:
|
||||
* Call `get_committee_assignment(state, next_epoch, validator_index)` when checking for next epoch assignments.
|
||||
* Find peers of the pubsub topic `committee_index{committee_index % ATTESTATION_SUBNET_COUNT}_beacon_attestation`.
|
||||
* 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"][committee_index % ATTESTATION_SUBNET_COUNT] == True`. Then validate that the peers are still persisted on the desired topic by requesting `GetMetaData` and checking the resulting `attnets` field.
|
||||
* Find peers of the pubsub topic `beacon_attestation_{compute_subnet_for_attestation(state, slot, committee_index)}`.
|
||||
* 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"][compute_subnet_for_attestation(state, slot, committee_index)] == 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.
|
||||
|
@ -425,18 +425,18 @@ def get_attestation_signature(state: BeaconState, attestation_data: AttestationD
|
|||
|
||||
#### Broadcast attestation
|
||||
|
||||
Finally, the validator broadcasts `attestation` to the associated attestation subnet -- the `beacon_attestation_{compute_subnet_for_attestation(state, attestation)}` pubsub topic.
|
||||
Finally, the validator broadcasts `attestation` to the associated attestation subnet -- the `beacon_attestation_{compute_subnet_for_attestation(state, attestation.data.slot, attestation.data.committee_index)}` pubsub topic.
|
||||
|
||||
```python
|
||||
def compute_subnet_for_attestation(state: BeaconState, attestation: Attestation) -> uint64:
|
||||
def compute_subnet_for_attestation(state: BeaconState, slot: Slot, committee_index: CommitteeIndex) -> uint64:
|
||||
"""
|
||||
Compute the correct subnet for an attestation for Phase 0.
|
||||
Note, this mimics expected Phase 1 behavior where attestations will be mapped to their shard subnet.
|
||||
"""
|
||||
slots_since_epoch_start = attestation.data.slot % SLOTS_PER_EPOCH
|
||||
committees_since_epoch_start = get_committee_count_at_slot(state, attestation.data.slot) * slots_since_epoch_start
|
||||
slots_since_epoch_start = slot % SLOTS_PER_EPOCH
|
||||
committees_since_epoch_start = get_committee_count_at_slot(state, slot) * slots_since_epoch_start
|
||||
|
||||
return (committees_since_epoch_start + attestation.data.index) % ATTESTATION_SUBNET_COUNT
|
||||
return (committees_since_epoch_start + committee_index) % ATTESTATION_SUBNET_COUNT
|
||||
```
|
||||
|
||||
### Attestation aggregation
|
||||
|
|
Loading…
Reference in New Issue