module

package
v0.15.3-patch.6 Latest Latest
Warning

This package is not in the latest version of its module.

 Go to latest Published: Apr 9, 2021 License: AGPL-3.0 Imports: 15 Imported by: 64

Documentation

Index

Constants

View Source 
const (
	ConsumeProgressVerificationBlockHeight = "ConsumeProgressVerificationBlockHeight"
	ConsumeProgressVerificationChunkIndex  = "ConsumeProgressVerificationChunkIndex"
)

Variables

This section is empty.

Functions

This section is empty.

Types

type AggregatingSigner 

type AggregatingSigner interface {
	AggregatingVerifier
	Sign(msg []byte) (crypto.Signature, error)
	Aggregate(sigs []crypto.Signature) (crypto.Signature, error)
}

AggregatingSigner is a signer that can sign a simple message and aggregate multiple signatures into a single aggregated signature.

type AggregatingVerifier 

type AggregatingVerifier interface {
	Verifier
	VerifyMany(msg []byte, sig crypto.Signature, keys []crypto.PublicKey) (bool, error)
}

AggregatingVerifier can verify a message against a signature from either a single key or many keys.

type ApprovalValidator added in v0.15.0 

type ApprovalValidator interface {
	Validate(approval *flow.ResultApproval) error
}

ApprovalValidator is used for validating result approvals received from verification nodes with respect to current protocol state. Returns the following: * nil - in case of success * sentinel engine.InvalidInputError when approval is invalid * sentinel engine.OutdatedInputError if the corresponding block has a finalized seal * sentinel engine.UnverifiableInputError if approval cannot be validation because of missing data * exception in case of any other error, usually this is not expected.

type BlockRequester 

type BlockRequester interface {

	// RequestBlock indicates that the given block should be queued for retrieval.
	RequestBlock(blockID flow.Identifier)

	// RequestHeight indicates that the given block height should be queued for retrieval.
	RequestHeight(height uint64)

	// Manually Prune requests
	Prune(final *flow.Header)
}

BlockRequester enables components to request particular blocks by ID from synchronization system.

type Builder 

type Builder interface {

	// BuildOn generates a new payload that is valid with respect to the parent
	// being built upon, with the view being provided by the consensus algorithm.
	// The builder stores the block and validates it against the protocol state
	// before returning it.
	//
	// NOTE: Since the block is stored within Builder, HotStuff MUST propose the
	// block once BuildOn succcessfully returns.
	BuildOn(parentID flow.Identifier, setter func(*flow.Header) error) (*flow.Header, error)
}

Builder represents an abstracted block construction module that can be used in more than one consensus algorithm. The resulting block is consistent within itself and can be wrapped with additional consensus information such as QCs.

type CacheMetrics 

type CacheMetrics interface {
	CacheEntries(resource string, entries uint)
	CacheHit(resource string)
	CacheMiss(resource string)
}

type ChunkAssigner 

type ChunkAssigner interface {
	// Assign generates the assignment
	// error returns:
	//  * NoValidChildBlockError indicates that no valid child block is known
	//    (which contains the block's source of randomness)
	//  * unexpected errors should be considered symptoms of internal bugs
	Assign(result *flow.ExecutionResult, blockID flow.Identifier) (*chmodels.Assignment, error)
}

ChunkAssigner presents an interface for assigning chunks to the verifier nodes

type ChunkVerifier 

type ChunkVerifier interface {
	// Verify verifies the given VerifiableChunk by executing it and checking the final state commitment
	// It returns a Spock Secret as a byte array, verification fault of the chunk, and an error.
	// Note: Verify should only be executed on non-system chunks. It returns an error if it is invoked on
	// system chunk.
	// TODO return challenges plus errors
	Verify(ch *verification.VerifiableChunkData) ([]byte, chmodels.ChunkFault, error)

	// VerifySystemChunk verifies a given VerifiableChunk corresponding to a system chunk.
	// by executing it and checking the final state commitment
	// It returns a Spock Secret as a byte array, verification fault of the chunk, and an error.
	// Note: Verify should only be executed on system chunks. It returns an error if it is invoked on
	// non-system chunks.
	SystemChunkVerify(ch *verification.VerifiableChunkData) ([]byte, chmodels.ChunkFault, error)
}

ChunkVerifier provides functionality to verify chunks

type CleanerMetrics 

type CleanerMetrics interface {
	RanGC(took time.Duration)
}

type ClusterRootQCVoter 

type ClusterRootQCVoter interface {

	// Vote handles the full procedure of generating a vote, submitting it to the
	// epoch smart contract, and verifying submission. Returns an error only if
	// there is a critical error that would make it impossible for the vote to be
	// submitted. Otherwise, exits when the vote has been successfully submitted.
	//
	// It is safe to run Vote multiple times within a single setup phase.
	Vote(context.Context, protocol.Epoch) error
}

ClusterRootQCVoter is responsible for submitting a vote to the cluster QC contract to coordinate generation of a valid root quorum certificate for the next epoch.

type ColdStuff 

type ColdStuff interface {
	HotStuff

	SubmitCommit(commit *model.Commit)
}

ColdStuff is the interface for accepting proposals, votes, and commits to the ColdStuff consensus algorithm.

NOTE: It re-uses as much of the HotStuff interface and models as possible to simplify swapping between the two.

type CollectionMetrics 

type CollectionMetrics interface {
	// TransactionIngested is called when a new transaction is ingested by the
	// node. It increments the total count of ingested transactions and starts
	// a tx->col span for the transaction.
	TransactionIngested(txID flow.Identifier)

	// ClusterBlockProposed is called when a new collection is proposed by us or
	// any other node in the cluster.
	ClusterBlockProposed(block *cluster.Block)

	// ClusterBlockFinalized is called when a collection is finalized.
	ClusterBlockFinalized(block *cluster.Block)
}

type ComplianceMetrics 

type ComplianceMetrics interface {
	FinalizedHeight(height uint64)
	SealedHeight(height uint64)
	BlockFinalized(*flow.Block)
	BlockSealed(*flow.Block)
	BlockProposalDuration(duration time.Duration)
}

type ConsensusMetrics 

type ConsensusMetrics interface {
	// StartCollectionToFinalized reports Metrics C1: Collection Received by CCL→ Collection Included in Finalized Block
	StartCollectionToFinalized(collectionID flow.Identifier)

	// FinishCollectionToFinalized reports Metrics C1: Collection Received by CCL→ Collection Included in Finalized Block
	FinishCollectionToFinalized(collectionID flow.Identifier)

	// StartBlockToSeal reports Metrics C4: Block Received by CCL → Block Seal in finalized block
	StartBlockToSeal(blockID flow.Identifier)

	// FinishBlockToSeal reports Metrics C4: Block Received by CCL → Block Seal in finalized block
	FinishBlockToSeal(blockID flow.Identifier)

	// EmergencySeal increments the number of seals that were created in emergency mode
	EmergencySeal()

	// OnReceiptProcessingDuration records the number of seconds spent processing a receipt
	OnReceiptProcessingDuration(duration time.Duration)

	// OnApprovalProcessingDuration records the number of seconds spent processing an approval
	OnApprovalProcessingDuration(duration time.Duration)

	// CheckSealingDuration records absolute time for the full sealing check by the consensus match engine
	CheckSealingDuration(duration time.Duration)
}

type Engine 

type Engine interface {
	ReadyDoneAware
	network.Engine
}

Engine is the interface all engines should implement in order to have a manageable lifecycle and recieve messages from the networking layer.

type EngineMetrics 

type EngineMetrics interface {
	MessageSent(engine string, message string)
	MessageReceived(engine string, message string)
	MessageHandled(engine string, messages string)
}

type EntriesFunc added in v0.14.6 

type EntriesFunc func() uint

type ExecutionMetrics 

type ExecutionMetrics interface {
	LedgerMetrics
	RuntimeMetrics
	ProviderMetrics
	WALMetrics

	// StartBlockReceivedToExecuted starts a span to trace the duration of a block
	// from being received for execution to execution being finished
	StartBlockReceivedToExecuted(blockID flow.Identifier)

	// FinishBlockReceivedToExecuted finishes a span to trace the duration of a block
	// from being received for execution to execution being finished
	FinishBlockReceivedToExecuted(blockID flow.Identifier)

	// ExecutionGasUsedPerBlock reports gas used per block
	ExecutionGasUsedPerBlock(gas uint64)

	// ExecutionStateReadsPerBlock reports number of state access/read operations per block
	ExecutionStateReadsPerBlock(reads uint64)

	// ExecutionStorageStateCommitment reports the storage size of a state commitment in bytes
	ExecutionStorageStateCommitment(bytes int64)

	// ExecutionLastExecutedBlockHeight reports last executed block height
	ExecutionLastExecutedBlockHeight(height uint64)

	// ExecutionTotalExecutedTransactions adds num to the total number of executed transactions
	ExecutionTotalExecutedTransactions(numExecuted int)

	// ExecutionCollectionRequestSent reports when a request for a collection is sent to a collection node
	ExecutionCollectionRequestSent()

	// Unused
	ExecutionCollectionRequestRetried()

	// ExecutionSync reports when the state syncing is triggered or stopped.
	ExecutionSync(syncing bool)
}

type Finalizer 

type Finalizer interface {

	// MakeValid will mark a block as having passed the consensus algorithm's
	// internal validation.
	MakeValid(blockID flow.Identifier) error

	// MakeFinal will declare a block and all of its ancestors as finalized, which
	// makes it an immutable part of the blockchain. Returning an error indicates
	// some fatal condition and will cause the finalization logic to terminate.
	MakeFinal(blockID flow.Identifier) error
}

Finalizer is used by the consensus algorithm to inform other components for (such as the protocol state) about validity of block headers and finalization of blocks.

Since we have two different protocol states: one for the main consensus, the other for the collection cluster consensus, the Finalizer interface allows the two different protocol states to provide different implementations for updating its state when a block has been validated or finalized.

Why MakeValid and MakeFinal need to return an error? Updating the protocol state should always succeed when the data is consistent. However, in case the protocol state is corrupted, error should be returned and the consensus algorithm should halt. So the error returned from MakeValid and MakeFinal is for the protocol state to report exceptions.

type HotStuff 

type HotStuff interface {
	ReadyDoneAware

	// SubmitProposal submits a new block proposal to the HotStuff event loop.
	// This method blocks until the proposal is accepted to the event queue.
	//
	// Block proposals must be submitted in order and only if they extend a
	// block already known to HotStuff core.
	SubmitProposal(proposal *flow.Header, parentView uint64)

	// SubmitVote submits a new vote to the HotStuff event loop.
	// This method blocks until the vote is accepted to the event queue.
	//
	// Votes may be submitted in any order.
	SubmitVote(originID flow.Identifier, blockID flow.Identifier, view uint64, sigData []byte)
}

HotStuff defines the interface to the core HotStuff algorithm. It includes a method to start the event loop, and utilities to submit block proposals and votes received from other replicas.

type HotStuffFollower 

type HotStuffFollower interface {
	ReadyDoneAware

	// SubmitProposal feeds a new block proposal into the HotStuffFollower.
	// This method blocks until the proposal is accepted to the event queue.
	//
	// Block proposals must be submitted in order, i.e. a proposal's parent must
	// have been previously processed by the HotStuffFollower.
	SubmitProposal(proposal *flow.Header, parentView uint64)
}

HotStuffFollower is run by non-consensus nodes to observe the block chain and make local determination about block finalization. While the process of reaching consensus (while guaranteeing its safety and liveness) is very intricate, the criteria to confirm that consensus has been reached are relatively straight forward. Each non-consensus node can simply observe the blockchain and determine locally which blocks have been finalized without requiring additional information from the consensus nodes.

Specifically, the HotStuffFollower informs other components within the node about finalization of blocks. It consumes block proposals broadcasted by the consensus node, verifies the block header and locally evaluates the finalization rules.

Notes:

  • HotStuffFollower does not handle disconnected blocks. Each block's parent must have been previously processed by the HotStuffFollower.
  • HotStuffFollower internally prunes blocks below the last finalized view. When receiving a block proposal, it might not have the proposal's parent anymore. Nevertheless, HotStuffFollower needs the parent's view, which must be supplied in addition to the proposal.

type HotstuffMetrics 

type HotstuffMetrics interface {
	// HotStuffBusyDuration reports Metrics C6 HotStuff Busy Duration
	HotStuffBusyDuration(duration time.Duration, event string)

	// HotStuffIdleDuration reports Metrics C6 HotStuff Idle Duration
	HotStuffIdleDuration(duration time.Duration)

	// HotStuffWaitDuration reports Metrics C6 HotStuff Idle Duration
	HotStuffWaitDuration(duration time.Duration, event string)

	// SetCurView reports Metrics C8: Current View
	SetCurView(view uint64)

	// SetQCView reports Metrics C9: View of Newest Known QC
	SetQCView(view uint64)

	// CountSkipped reports the number of times we skipped ahead.
	CountSkipped()

	// CountTimeout reports the number of times we timed out.
	CountTimeout()

	// SetTimeout sets the current timeout duration
	SetTimeout(duration time.Duration)

	// CommitteeProcessingDuration measures the time which the HotStuff's core logic
	// spends in the hotstuff.Committee component, i.e. the time determining consensus
	// committee relations.
	CommitteeProcessingDuration(duration time.Duration)

	// SignerProcessingDuration measures the time which the HotStuff's core logic
	// spends in the hotstuff.Signer component, i.e. the with crypto-related operations.
	SignerProcessingDuration(duration time.Duration)

	// ValidatorProcessingDuration measures the time which the HotStuff's core logic
	// spends in the hotstuff.Validator component, i.e. the with verifying
	// consensus messages.
	ValidatorProcessingDuration(duration time.Duration)

	// PayloadProductionDuration measures the time which the HotStuff's core logic
	// spends in the module.Builder component, i.e. the with generating block payloads.
	PayloadProductionDuration(duration time.Duration)
}

type Job added in v0.15.0 

type Job interface {
	// each job has a unique ID for deduplication
	ID() JobID
}

type JobConsumer added in v0.15.0 

type JobConsumer interface {
	NewJobListener

	// Start starts processing jobs from a job queue. If this is the first time, a processed index
	// will be initialized in the storage. If it fails to initialize, an error will be returned
	Start(defaultIndex uint64) error

	// Stop gracefully stops the consumer from reading new jobs from the job queue. It does not stop
	// the existing worker finishing their jobs
	// It blocks until the existing worker finish processing the job
	Stop()

	// NotifyJobIsDone let the workers notify consumer that a job has been finished, so that the consumer
	// can check if there is new job could be read from storage and give to a worker for processing
	NotifyJobIsDone(JobID)
}

JobConsumer consumes jobs from a job queue, and it remembers which job it has processed, and is able to resume processing from the next.

type JobID added in v0.15.0 

type JobID string

JobID is a unique ID of the job.

type JobQueue added in v0.15.0 

type JobQueue interface {
	// Add a job to the job queue
	Add(job Job) error
}

type Jobs added in v0.15.0 

type Jobs interface {
	AtIndex(index uint64) (Job, error)

	// Head returns the index of the last job
	Head() (uint64, error)
}

Jobs is the reader for an ordered job queue. Job can be fetched by the index, which start from 0

type LedgerMetrics 

type LedgerMetrics interface {
	// ForestApproxMemorySize records approximate memory usage of forest (all in-memory trees)
	ForestApproxMemorySize(bytes uint64)

	// ForestNumberOfTrees current number of trees in a forest (in memory)
	ForestNumberOfTrees(number uint64)

	// LatestTrieRegCount records the number of unique register allocated (the latest created trie)
	LatestTrieRegCount(number uint64)

	// LatestTrieRegCountDiff records the difference between the number of unique register allocated of the latest created trie and parent trie
	LatestTrieRegCountDiff(number uint64)

	// LatestTrieMaxDepth records the maximum depth of the last created trie
	LatestTrieMaxDepth(number uint64)

	// LatestTrieMaxDepthDiff records the difference between the max depth of the latest created trie and parent trie
	LatestTrieMaxDepthDiff(number uint64)

	// UpdateCount increase a counter of performed updates
	UpdateCount()

	// ProofSize records a proof size
	ProofSize(bytes uint32)

	// UpdateValuesNumber accumulates number of updated values
	UpdateValuesNumber(number uint64)

	// UpdateValuesSize total size (in bytes) of updates values
	UpdateValuesSize(byte uint64)

	// UpdateDuration records absolute time for the update of a trie
	UpdateDuration(duration time.Duration)

	// UpdateDurationPerItem records update time for single value (total duration / number of updated values)
	UpdateDurationPerItem(duration time.Duration)

	// ReadValuesNumber accumulates number of read values
	ReadValuesNumber(number uint64)

	// ReadValuesSize total size (in bytes) of read values
	ReadValuesSize(byte uint64)

	// ReadDuration records absolute time for the read from a trie
	ReadDuration(duration time.Duration)

	// ReadDurationPerItem records read time for single value (total duration / number of read values)
	ReadDurationPerItem(duration time.Duration)
}

LedgerMetrics provides an interface to record Ledger Storage metrics. Ledger storage is non-linear (fork-aware) so certain metrics are averaged and computed before emitting for better visibility

type Local 

type Local interface {

	// NodeID returns the node ID of the local node.
	NodeID() flow.Identifier

	// Address returns the (listen) address of the local node.
	Address() string

	// Sign provides a signature oracle that given a message and hasher, it
	// generates and returns a signature over the message using the node's private key
	// as well as the input hasher
	Sign([]byte, hash.Hasher) (crypto.Signature, error)

	// NotMeFilter returns handy not-me filter for searching identity
	NotMeFilter() flow.IdentityFilter

	// SignFunc provides a signature oracle that given a message, a hasher, and a signing function, it
	// generates and returns a signature over the message using the node's private key
	// as well as the input hasher by invoking the given signing function. The overall idea of this function
	// is to not expose the private key to the caller.
	SignFunc([]byte, hash.Hasher, func(crypto.PrivateKey, []byte, hash.Hasher) (crypto.Signature,
		error)) (crypto.Signature, error)
}

Local encapsulates the stable local node information.

type MempoolMetrics 

type MempoolMetrics interface {
	MempoolEntries(resource string, entries uint)
	Register(resource string, entriesFunc EntriesFunc) error
}

type Merger 

type Merger interface {
	// Join concatenates the provided signatures. The merger has an internal notion
	// of the expected length of each signature. It returns the sentinel error
	// `verification.ErrInvalidFormat` if one of the signatures does not conform
	// to the expected byte length.
	Join(sig1, sig2 crypto.Signature) ([]byte, error)
	// Split separates the concatenated signature into its two components. The
	// merger has an internal notion of the expected byte length of each
	// signature. It returns the sentinel error `verification.ErrInvalidFormat`
	// if either signatures does not conform to the expected length.
	Split(combined []byte) (crypto.Signature, crypto.Signature, error)
}

Merger is responsible for combining two signatures, but it must be done in a cryptographically unaware way (agnostic of the byte structure of the signatures). However, the merger has an internal notion of the expected length of each signature and errors in case the inputs have incompatible length.

type Network 

type Network interface {

	// Register will subscribe to the channel with the given engine and
	// the engine will be notified with incoming messages on the channel.
	// The returned Conduit can be used to send messages to engines on other nodes subscribed to the same channel
	// On a single node, only one engine can be subscribed to a channel at any given time.
	Register(channel network.Channel, engine network.Engine) (network.Conduit, error)
}

Network represents the network layer of the node. It allows processes that work across the peer-to-peer network to register themselves as an engine with a unique engine ID. The returned conduit allows the process to communicate to the same engine on other nodes across the network in a network-agnostic way.

type NetworkMetrics 

type NetworkMetrics interface {
	// NetworkMessageSent size in bytes and count of the network message sent
	NetworkMessageSent(sizeBytes int, topic string, messageType string)

	// NetworkMessageReceived size in bytes and count of the network message received
	NetworkMessageReceived(sizeBytes int, topic string, messageType string)

	// NetworkDuplicateMessagesDropped counts number of messages dropped due to duplicate detection
	NetworkDuplicateMessagesDropped(topic string, messageType string)

	// Message receive queue metrics
	// MessageAdded increments the metric tracking the number of messages in the queue with the given priority
	MessageAdded(priority int)

	// MessageRemoved decrements the metric tracking the number of messages in the queue with the given priority
	MessageRemoved(priority int)

	// QueueDuration tracks the time spent by a message with the given priority in the queue
	QueueDuration(duration time.Duration, priority int)

	// InboundProcessDuration tracks the time a queue worker blocked by an engine for processing an incoming message on specified topic (i.e., channel).
	InboundProcessDuration(topic string, duration time.Duration)

	// OutboundConnections updates the metric tracking the number of outbound connections of this node
	OutboundConnections(connectionCount uint)

	// InboundConnections updates the metric tracking the number of inbound connections of this node
	InboundConnections(connectionCount uint)
}

Network Metrics

type NewJobListener added in v0.15.0 

type NewJobListener interface {
	// Check let the producer notify the consumer that a new job has been added, so that the consumer
	// can check if there is worker available to process that job.
	Check()
}

type PendingBlockBuffer 

type PendingBlockBuffer interface {
	Add(originID flow.Identifier, proposal *messages.BlockProposal) bool

	ByID(blockID flow.Identifier) (*flow.PendingBlock, bool)

	ByParentID(parentID flow.Identifier) ([]*flow.PendingBlock, bool)

	DropForParent(parentID flow.Identifier)

	PruneByHeight(height uint64)

	Size() uint
}

PendingBlockBuffer defines an interface for a cache of pending blocks that cannot yet be processed because they do not connect to the rest of the chain state. They are indexed by parent ID to enable processing all of a parent's children once the parent is received.

type PendingClusterBlockBuffer 

type PendingClusterBlockBuffer interface {
	Add(originID flow.Identifier, proposal *messages.ClusterBlockProposal) bool

	ByID(blockID flow.Identifier) (*cluster.PendingBlock, bool)

	ByParentID(parentID flow.Identifier) ([]*cluster.PendingBlock, bool)

	DropForParent(parentID flow.Identifier)

	PruneByHeight(height uint64)

	Size() uint
}

PendingClusterBlockBuffer is the same thing as PendingBlockBuffer, but for collection node cluster consensus.

type PingMetrics 

type PingMetrics interface {
	// NodeReachable tracks the round trip time in milliseconds taken to ping a node
	// The nodeInfo provides additional information about the node such as the name of the node operator
	NodeReachable(node *flow.Identity, nodeInfo string, rtt time.Duration)
}

type ProcessingNotifier added in v0.15.0 

type ProcessingNotifier interface {
	Notify(entityID flow.Identifier)
}

ProcessingNotifier is for the worker's underneath engine to report an entity has been processed without knowing the job queue. It is a callback so that the worker can convert the entity id into a job id, and notify the consumer about a finished job.

At the current version, entities used in this interface are chunks and blocks ids.

type ProviderMetrics 

type ProviderMetrics interface {
	// ChunkDataPackRequested is executed every time a chunk data pack request is arrived at execution node.
	// It increases the request counter by one.
	ChunkDataPackRequested()
}

type QCContractClient 

type QCContractClient interface {

	// SubmitVote submits the given vote to the cluster QC aggregator smart
	// contract. This function returns only once the transaction has been
	// processed by the network. An error is returned if the transaction has
	// failed and should be re-submitted.
	SubmitVote(ctx context.Context, vote *model.Vote) error

	// Voted returns true if we have successfully submitted a vote to the
	// cluster QC aggregator smart contract for the current epoch.
	Voted(ctx context.Context) (bool, error)
}

QCContractClient enables interacting with the cluster QC aggregator smart contract. This contract is deployed to the service account as part of a collection of smart contracts that facilitate and manage epoch transitions.

type ReadyDoneAware 

type ReadyDoneAware interface {
	Ready() <-chan struct{}
	Done() <-chan struct{}
}

ReadyDoneAware provides easy interface to wait for module startup and shutdown

type ReceiptValidator added in v0.14.0 

type ReceiptValidator interface {

	// Validate verifies that the ExecutionReceipt satisfies
	// the following conditions:
	// 	* is from Execution node with positive weight
	//	* has valid signature
	//	* chunks are in correct format
	// 	* execution result has a valid parent and satisfies the subgraph check
	// Returns nil if all checks passed successfully.
	// Expected errors during normal operations:
	// * engine.InvalidInputError
	//   if receipt violates protocol condition
	// * engine.UnverifiableInputError
	//   if receipt's parent result is unknown
	Validate(receipts *flow.ExecutionReceipt) error

	// ValidatePayload verifies the ExecutionReceipts and ExecutionResults
	// in the payload for compliance with the protocol:
	// Receipts:
	// 	* are from Execution node with positive weight
	//	* have valid signature
	//	* chunks are in correct format
	//  * no duplicates in fork
	// Results:
	// 	* have valid parents and satisfy the subgraph check
	//  * extend the execution tree, where the tree root is the latest
	//    finalized block and only results from this fork are included
	//  * no duplicates in fork
	// Expected errors during normal operations:
	// * engine.InvalidInputError
	//   if some receipts in the candidate block violate protocol condition
	// * engine.UnverifiableInputError
	//   if for some of the receipts, their respective parent result is unknown
	ValidatePayload(candidate *flow.Block) error
}

ReceiptValidator is an interface which is used for validating receipts with respect to current protocol state.

type Requester 

type Requester interface {
	// EntityByID will request an entity through the request engine backing
	// the interface. The additional selector will be applied to the subset
	// of valid providers for the entity and allows finer-grained control
	// over which providers to request a given entity from. Use `filter.Any`
	// if no additional restrictions are required. Data integrity of response
	// will be checked upon arrival. This function should be used for requesting
	// entites by their IDs.
	EntityByID(entityID flow.Identifier, selector flow.IdentityFilter)

	// Query will request data through the request engine backing the interface.
	// The additional selector will be applied to the subset
	// of valid providers for the data and allows finer-grained control
	// over which providers to request data from. Doesn't perform integrity check
	// can be used to get entities without knowing their ID.
	Query(key flow.Identifier, selector flow.IdentityFilter)

	// Force will force the dispatcher to send all possible batches immediately.
	// It can be used in cases where responsiveness is of utmost importance, at
	// the cost of additional network messages.
	Force()
}

type RuntimeMetrics 

type RuntimeMetrics interface {
	// TransactionParsed reports the time spent parsing a single transaction
	TransactionParsed(dur time.Duration)

	// TransactionChecked reports the time spent checking a single transaction
	TransactionChecked(dur time.Duration)

	// TransactionInterpreted reports the time spent interpreting a single transaction
	TransactionInterpreted(dur time.Duration)
}

type SealValidator added in v0.14.0 

type SealValidator interface {
	Validate(candidate *flow.Block) (*flow.Seal, error)
}

SealValidator checks seals with respect to current protocol state. Accepts `candidate` block with seals that needs to be verified for protocol state validity. Returns the following values: * last seal in `candidate` block - in case of success * engine.InvalidInputError - in case if `candidate` block violates protocol state. * exception in case of any other error, usually this is not expected. PREREQUISITE: The SealValidator can only process blocks which are attached to the main chain (without any missing ancestors). This is the case because:

  • the Seal validator walks the chain backwards and requires the relevant ancestors to be known and validated
  • the storage.Seals only holds seals for block that are attached to the main chain.

type Signer 

type Signer interface {
	Verifier
	Sign(msg []byte) (crypto.Signature, error)
}

Signer is a simple cryptographic signer that can sign a simple message to generate a signature, and verify the signature against the message.

type SyncCore 

type SyncCore interface {

	// HandleBlock handles receiving a new block. It returns true if the block
	// should be passed along to the rest of the system for processing, or false
	// if it should be discarded.
	HandleBlock(header *flow.Header) bool

	// HandleHeight handles receiving a new highest finalized height from another node.
	HandleHeight(final *flow.Header, height uint64)

	// ScanPending scans all pending block statuses for blocks that should be
	// requested. It apportions requestable items into range and batch requests
	// according to configured maximums, giving precedence to range requests.
	ScanPending(final *flow.Header) ([]flow.Range, []flow.Batch)

	// WithinTolerance returns whether or not the given height is within configured
	// height tolerance, wrt the given local finalized header.
	WithinTolerance(final *flow.Header, height uint64) bool

	// RangeRequested updates sync state after a range is requested.
	RangeRequested(ran flow.Range)

	// BatchRequested updates sync state after a batch is requested.
	BatchRequested(batch flow.Batch)
}

SyncCore represents state management for chain state synchronization.

type ThresholdSigner 

type ThresholdSigner interface {
	ThresholdVerifier
	Sign(msg []byte) (crypto.Signature, error)
	Reconstruct(size uint, shares []crypto.Signature, indices []uint) (crypto.Signature, error)
}

ThresholdSigner is a signer that can sign a message to generate a signature share and construct a threshold signature from the given shares.

type ThresholdVerifier 

type ThresholdVerifier interface {
	Verifier
	VerifyThreshold(msg []byte, sig crypto.Signature, key crypto.PublicKey) (bool, error)
}

ThresholdVerifier can verify a message against a signature share from a single key or a threshold signature against many keys.

type Tracer 

type Tracer interface {
	ReadyDoneAware
	StartSpan(entity flow.Identifier, spanName trace.SpanName, opts ...opentracing.StartSpanOption) opentracing.Span
	FinishSpan(entity flow.Identifier, spanName trace.SpanName)
	GetSpan(entity flow.Identifier, spanName trace.SpanName) (opentracing.Span, bool)

	StartSpanFromContext(
		ctx context.Context,
		operationName trace.SpanName,
		opts ...opentracing.StartSpanOption,
	) (opentracing.Span, context.Context)

	StartSpanFromParent(
		span opentracing.Span,
		operationName trace.SpanName,
		opts ...opentracing.StartSpanOption,
	) opentracing.Span

	// RecordSpanFromParent records an span at finish time
	// start time will be computed by reducing time.Now() - duration
	RecordSpanFromParent(
		span opentracing.Span,
		operationName trace.SpanName,
		duration time.Duration,
		logs []opentracing.LogRecord,
		opts ...opentracing.StartSpanOption,
	)

	// WithSpanFromContext encapsulates executing a function within an span, i.e., it starts a span with the specified SpanName from the context,
	// executes the function f, and finishes the span once the function returns.
	WithSpanFromContext(ctx context.Context,
		operationName trace.SpanName,
		f func(),
		opts ...opentracing.StartSpanOption)
}

Tracer interface for tracers in flow. Uses open tracing span definitions

type TransactionMetrics 

type TransactionMetrics interface {
	// TransactionReceived starts tracking of transaction execution/finalization/sealing
	TransactionReceived(txID flow.Identifier, when time.Time)

	// TransactionFinalized reports the time spent between the transaction being received and finalized. Reporting only
	// works if the transaction was earlier added as received.
	TransactionFinalized(txID flow.Identifier, when time.Time)

	// TransactionExecuted reports the time spent between the transaction being received and executed. Reporting only
	// works if the transaction was earlier added as received.
	TransactionExecuted(txID flow.Identifier, when time.Time)

	// TransactionExpired tracks number of expired transactions
	TransactionExpired(txID flow.Identifier)

	// TransactionSubmissionFailed should be called whenever we try to submit a transaction and it fails
	TransactionSubmissionFailed()
}

type VerificationMetrics 

type VerificationMetrics interface {
	// Finder Engine
	//
	// OnExecutionReceiptReceived is called whenever a new execution receipt arrives
	// at Finder engine. It increments total number of received receipts.
	OnExecutionReceiptReceived()
	// OnExecutionResultSent is called whenever a new execution result is sent by
	// Finder engine to the match engine. It increments total number of sent execution results.
	OnExecutionResultSent()

	// Match Engine
	//
	// OnExecutionResultReceived is called whenever a new execution result is successfully received
	// by Match engine from Finder engine.
	// It increments the total number of received execution results.
	OnExecutionResultReceived()
	// OnVerifiableChunkSent is called on a successful submission of matched chunk
	// by Match engine to Verifier engine.
	// It increments the total number of chunks matched by match engine.
	OnVerifiableChunkSent()

	// OnChunkDataPackReceived is called on a receiving a chunk data pack by Match engine
	// It increments the total number of chunk data packs received.
	OnChunkDataPackReceived()

	// OnChunkDataPackRequested is called on requesting a chunk data pack by Match engine
	// It increments the total number of chunk data packs requested.
	OnChunkDataPackRequested()

	// Verifier Engine
	//
	// OnVerifiableChunkReceived is called whenever a verifiable chunk is received by Verifier engine
	// from Match engine.It increments the total number of sent verifiable chunks.
	OnVerifiableChunkReceived()
	// OnResultApproval is called whenever a result approval for is emitted to consensus nodes.
	// It increases the total number of result approvals.
	OnResultApproval()

	// OnFinalizedBlockReceived is called whenever a finalized block arrives at the assigner engine.
	// It increments the total number of finalized blocks.
	//
	//
	// Note: it assumes blocks are coming to assigner engine in strictly increasing order of their height.
	OnAssignerProcessFinalizedBlock(height uint64)

	// OnChunksAssigned is called whenever chunks assigned to this verification node by applying chunk assignment on an
	// execution result.
	// It increases the total number of assigned chunks by the input.
	OnChunksAssigned(chunks int)

	// OnChunkProcessed is called whenever a chunk is pushed to the chunks queue by the assigner engine.
	// It increments the total number of sent chunks.
	OnChunkProcessed()
}

type Verifier 

type Verifier interface {
	Verify(msg []byte, sig crypto.Signature, key crypto.PublicKey) (bool, error)
}

Verifier is responsible for generating a signature on the given message.

type WALMetrics added in v0.15.4 

type WALMetrics interface {
	// DiskSize records the amount of disk space used by the storage (in bytes)
	DiskSize(uint64)
}

Source Files

View all Source files

Directories

Path Synopsis
builder
collection
consensus
finalizer
collection
consensus
mock
Package ingress implements accepting transactions into the system.
 Package ingress implements accepting transactions into the system.
(c) 2019 Dapper Labs - ALL RIGHTS RESERVED
 (c) 2019 Dapper Labs - ALL RIGHTS RESERVED
consensus
consensus/mock
ejectors
entity
epochs
mock
model
queue
stdmap
(c) 2019 Dapper Labs - ALL RIGHTS RESERVED
 (c) 2019 Dapper Labs - ALL RIGHTS RESERVED
example
example/collection
example/consensus
example/execution
example/verification
hotstuff
Package mocks is a generated GoMock package.
 Package mocks is a generated GoMock package.

Jump to

Keyboard shortcuts

? : This menu
/ : Search site
f or F : Jump to
y or Y : Canonical URL
go.dev uses cookies from Google to deliver and enhance the quality of its services and to analyze traffic. Learn more.