lnd.xprv/routing/pathfind.go

911 lines
30 KiB
Go
Raw Normal View History

package routing
routing: rewrite package to conform to BOLT07 and factor in fees+timelocks This commit overhauls the routing package significantly to simplify the code, conform to the rest of the coding style within the package, and observe the new authenticated gossiping scheme outlined in BOLT07. As a major step towards a more realistic path finding algorithm, fees are properly calculated and observed during path finding. If a path has sufficient capacity _before_ fees are applied, but afterwards the finalized route would exceed the capacity of a single link, the route is marked as invalid. Currently a naive weighting algorithm is used which only factors in the time-lock delta at each hop, thereby optimizing for the lowest time lock. Fee calculation also isn’t finalized since we aren’t yet using milli-satoshi throughout the daemon. The final TODO item within the PR is to properly perform a multi-path search and rank the results based on a summation heuristic rather than just return the first (out of many) route found. On the server side, once nodes are initially connected to the daemon, our routing table will be synced with the peer’s using a naive “just send everything scheme” to hold us over until I spec out some a efficient graph reconciliation protocol. Additionally, the routing table is now pruned by the channel router itself once new blocks arrive rather than depending on peers to tell us when a channel flaps or is closed. Finally, the validation of peer announcements aren’t yet fully implemented as they’ll be implemented within the pending discovery package that was blocking on the completion of this package. Most off the routing message processing will be moved out of this package and into the discovery package where full validation will be carried out.
2016-12-27 08:20:26 +03:00
import (
"container/heap"
"errors"
"fmt"
"math"
2019-09-06 09:56:59 +03:00
"time"
sphinx "github.com/lightningnetwork/lightning-onion"
routing: rewrite package to conform to BOLT07 and factor in fees+timelocks This commit overhauls the routing package significantly to simplify the code, conform to the rest of the coding style within the package, and observe the new authenticated gossiping scheme outlined in BOLT07. As a major step towards a more realistic path finding algorithm, fees are properly calculated and observed during path finding. If a path has sufficient capacity _before_ fees are applied, but afterwards the finalized route would exceed the capacity of a single link, the route is marked as invalid. Currently a naive weighting algorithm is used which only factors in the time-lock delta at each hop, thereby optimizing for the lowest time lock. Fee calculation also isn’t finalized since we aren’t yet using milli-satoshi throughout the daemon. The final TODO item within the PR is to properly perform a multi-path search and rank the results based on a summation heuristic rather than just return the first (out of many) route found. On the server side, once nodes are initially connected to the daemon, our routing table will be synced with the peer’s using a naive “just send everything scheme” to hold us over until I spec out some a efficient graph reconciliation protocol. Additionally, the routing table is now pruned by the channel router itself once new blocks arrive rather than depending on peers to tell us when a channel flaps or is closed. Finally, the validation of peer announcements aren’t yet fully implemented as they’ll be implemented within the pending discovery package that was blocking on the completion of this package. Most off the routing message processing will be moved out of this package and into the discovery package where full validation will be carried out.
2016-12-27 08:20:26 +03:00
"github.com/lightningnetwork/lnd/channeldb"
"github.com/lightningnetwork/lnd/feature"
"github.com/lightningnetwork/lnd/lnwire"
"github.com/lightningnetwork/lnd/record"
"github.com/lightningnetwork/lnd/routing/route"
routing: rewrite package to conform to BOLT07 and factor in fees+timelocks This commit overhauls the routing package significantly to simplify the code, conform to the rest of the coding style within the package, and observe the new authenticated gossiping scheme outlined in BOLT07. As a major step towards a more realistic path finding algorithm, fees are properly calculated and observed during path finding. If a path has sufficient capacity _before_ fees are applied, but afterwards the finalized route would exceed the capacity of a single link, the route is marked as invalid. Currently a naive weighting algorithm is used which only factors in the time-lock delta at each hop, thereby optimizing for the lowest time lock. Fee calculation also isn’t finalized since we aren’t yet using milli-satoshi throughout the daemon. The final TODO item within the PR is to properly perform a multi-path search and rank the results based on a summation heuristic rather than just return the first (out of many) route found. On the server side, once nodes are initially connected to the daemon, our routing table will be synced with the peer’s using a naive “just send everything scheme” to hold us over until I spec out some a efficient graph reconciliation protocol. Additionally, the routing table is now pruned by the channel router itself once new blocks arrive rather than depending on peers to tell us when a channel flaps or is closed. Finally, the validation of peer announcements aren’t yet fully implemented as they’ll be implemented within the pending discovery package that was blocking on the completion of this package. Most off the routing message processing will be moved out of this package and into the discovery package where full validation will be carried out.
2016-12-27 08:20:26 +03:00
)
const (
// infinity is used as a starting distance in our shortest path search.
infinity = math.MaxInt64
// RiskFactorBillionths controls the influence of time lock delta
// of a channel on route selection. It is expressed as billionths
// of msat per msat sent through the channel per time lock delta
// block. See edgeWeight function below for more details.
// The chosen value is based on the previous incorrect weight function
// 1 + timelock + fee * fee. In this function, the fee penalty
// diminishes the time lock penalty for all but the smallest amounts.
// To not change the behaviour of path finding too drastically, a
// relatively small value is chosen which is still big enough to give
// some effect with smaller time lock values. The value may need
// tweaking and/or be made configurable in the future.
RiskFactorBillionths = 15
// estimatedNodeCount is used to preallocate the path finding structures
// to avoid resizing and copies. It should be number on the same order as
// the number of active nodes in the network.
estimatedNodeCount = 10000
routing: rewrite package to conform to BOLT07 and factor in fees+timelocks This commit overhauls the routing package significantly to simplify the code, conform to the rest of the coding style within the package, and observe the new authenticated gossiping scheme outlined in BOLT07. As a major step towards a more realistic path finding algorithm, fees are properly calculated and observed during path finding. If a path has sufficient capacity _before_ fees are applied, but afterwards the finalized route would exceed the capacity of a single link, the route is marked as invalid. Currently a naive weighting algorithm is used which only factors in the time-lock delta at each hop, thereby optimizing for the lowest time lock. Fee calculation also isn’t finalized since we aren’t yet using milli-satoshi throughout the daemon. The final TODO item within the PR is to properly perform a multi-path search and rank the results based on a summation heuristic rather than just return the first (out of many) route found. On the server side, once nodes are initially connected to the daemon, our routing table will be synced with the peer’s using a naive “just send everything scheme” to hold us over until I spec out some a efficient graph reconciliation protocol. Additionally, the routing table is now pruned by the channel router itself once new blocks arrive rather than depending on peers to tell us when a channel flaps or is closed. Finally, the validation of peer announcements aren’t yet fully implemented as they’ll be implemented within the pending discovery package that was blocking on the completion of this package. Most off the routing message processing will be moved out of this package and into the discovery package where full validation will be carried out.
2016-12-27 08:20:26 +03:00
)
// pathFinder defines the interface of a path finding algorithm.
type pathFinder = func(g *graphParams, r *RestrictParams,
cfg *PathFindingConfig, source, target route.Vertex,
amt lnwire.MilliSatoshi, finalHtlcExpiry int32) (
[]*channeldb.ChannelEdgePolicy, error)
var (
// DefaultPaymentAttemptPenalty is the virtual cost in path finding weight
// units of executing a payment attempt that fails. It is used to trade
// off potentially better routes against their probability of
// succeeding.
DefaultPaymentAttemptPenalty = lnwire.NewMSatFromSatoshis(100)
// DefaultMinRouteProbability is the default minimum probability for routes
// returned from findPath.
DefaultMinRouteProbability = float64(0.01)
// DefaultAprioriHopProbability is the default a priori probability for
// a hop.
DefaultAprioriHopProbability = float64(0.6)
)
2018-06-04 23:10:05 +03:00
// edgePolicyWithSource is a helper struct to keep track of the source node
// of a channel edge. ChannelEdgePolicy only contains to destination node
// of the edge.
type edgePolicyWithSource struct {
sourceNode route.Vertex
2018-06-04 23:10:05 +03:00
edge *channeldb.ChannelEdgePolicy
}
// finalHopParams encapsulates various parameters for route construction that
// apply to the final hop in a route. These features include basic payment data
// such as amounts and cltvs, as well as more complex features like destination
// custom records and payment address.
type finalHopParams struct {
amt lnwire.MilliSatoshi
cltvDelta uint16
records record.CustomSet
paymentAddr *[32]byte
}
// newRoute constructs a route using the provided path and final hop constraints.
// Any destination specific fields from the final hop params will be attached
// assuming the destination's feature vector signals support, otherwise this
// method will fail. If the route is too long, or the selected path cannot
// support the fully payment including fees, then a non-nil error is returned.
//
// NOTE: The passed slice of ChannelHops MUST be sorted in forward order: from
// the source to the target node of the path finding attempt. It is assumed that
// any feature vectors on all hops have been validated for transitive
// dependencies.
func newRoute(sourceVertex route.Vertex,
pathEdges []*channeldb.ChannelEdgePolicy, currentHeight uint32,
finalHop finalHopParams) (*route.Route, error) {
2018-10-24 03:16:39 +03:00
var (
hops []*route.Hop
routing: rewrite package to conform to BOLT07 and factor in fees+timelocks This commit overhauls the routing package significantly to simplify the code, conform to the rest of the coding style within the package, and observe the new authenticated gossiping scheme outlined in BOLT07. As a major step towards a more realistic path finding algorithm, fees are properly calculated and observed during path finding. If a path has sufficient capacity _before_ fees are applied, but afterwards the finalized route would exceed the capacity of a single link, the route is marked as invalid. Currently a naive weighting algorithm is used which only factors in the time-lock delta at each hop, thereby optimizing for the lowest time lock. Fee calculation also isn’t finalized since we aren’t yet using milli-satoshi throughout the daemon. The final TODO item within the PR is to properly perform a multi-path search and rank the results based on a summation heuristic rather than just return the first (out of many) route found. On the server side, once nodes are initially connected to the daemon, our routing table will be synced with the peer’s using a naive “just send everything scheme” to hold us over until I spec out some a efficient graph reconciliation protocol. Additionally, the routing table is now pruned by the channel router itself once new blocks arrive rather than depending on peers to tell us when a channel flaps or is closed. Finally, the validation of peer announcements aren’t yet fully implemented as they’ll be implemented within the pending discovery package that was blocking on the completion of this package. Most off the routing message processing will be moved out of this package and into the discovery package where full validation will be carried out.
2016-12-27 08:20:26 +03:00
2018-10-24 03:16:39 +03:00
// totalTimeLock will accumulate the cumulative time lock
// across the entire route. This value represents how long the
// sender will need to wait in the *worst* case.
totalTimeLock = currentHeight
2018-10-24 03:16:39 +03:00
// nextIncomingAmount is the amount that will need to flow into
// the *next* hop. Since we're going to be walking the route
// backwards below, this next hop gets closer and closer to the
// sender of the payment.
nextIncomingAmount lnwire.MilliSatoshi
)
2018-10-24 03:16:39 +03:00
pathLength := len(pathEdges)
for i := pathLength - 1; i >= 0; i-- {
// Now we'll start to calculate the items within the per-hop
2018-10-24 03:16:39 +03:00
// payload for the hop this edge is leading to.
edge := pathEdges[i]
// We'll calculate the amounts, timelocks, and fees for each hop
// in the route. The base case is the final hop which includes
// their amount and timelocks. These values will accumulate
// contributions from the preceding hops back to the sender as
// we compute the route in reverse.
var (
amtToForward lnwire.MilliSatoshi
fee lnwire.MilliSatoshi
outgoingTimeLock uint32
tlvPayload bool
customRecords record.CustomSet
mpp *record.MPP
)
// Define a helper function that checks this edge's feature
// vector for support for a given feature. We assume at this
// point that the feature vectors transitive dependencies have
// been validated.
supports := edge.Node.Features.HasFeature
// We start by assuming the node doesn't support TLV. We'll now
// inspect the node's feature vector to see if we can promote
// the hop. We assume already that the feature vector's
// transitive dependencies have already been validated by path
// finding or some other means.
tlvPayload = supports(lnwire.TLVOnionPayloadOptional)
if i == len(pathEdges)-1 {
// If this is the last hop, then the hop payload will
// contain the exact amount. In BOLT #4: Onion Routing
// Protocol / "Payload for the Last Node", this is
// detailed.
amtToForward = finalHop.amt
// Fee is not part of the hop payload, but only used for
// reporting through RPC. Set to zero for the final hop.
fee = lnwire.MilliSatoshi(0)
// As this is the last hop, we'll use the specified
// final CLTV delta value instead of the value from the
// last link in the route.
totalTimeLock += uint32(finalHop.cltvDelta)
outgoingTimeLock = totalTimeLock
// Attach any custom records to the final hop if the
// receiver supports TLV.
if !tlvPayload && finalHop.records != nil {
return nil, errors.New("cannot attach " +
"custom records")
}
customRecords = finalHop.records
// If we're attaching a payment addr but the receiver
// doesn't support both TLV and payment addrs, fail.
payAddr := supports(lnwire.PaymentAddrOptional)
if !payAddr && finalHop.paymentAddr != nil {
return nil, errors.New("cannot attach " +
"payment addr")
}
// Otherwise attach the mpp record if it exists.
if finalHop.paymentAddr != nil {
mpp = record.NewMPP(
finalHop.amt, *finalHop.paymentAddr,
)
}
} else {
2018-07-31 10:19:49 +03:00
// The amount that the current hop needs to forward is
// equal to the incoming amount of the next hop.
amtToForward = nextIncomingAmount
// The fee that needs to be paid to the current hop is
2018-07-31 10:19:49 +03:00
// based on the amount that this hop needs to forward
// and its policy for the outgoing channel. This policy
// is stored as part of the incoming channel of
// the next hop.
2019-01-24 23:14:05 +03:00
fee = pathEdges[i+1].ComputeFee(amtToForward)
routing: rewrite package to conform to BOLT07 and factor in fees+timelocks This commit overhauls the routing package significantly to simplify the code, conform to the rest of the coding style within the package, and observe the new authenticated gossiping scheme outlined in BOLT07. As a major step towards a more realistic path finding algorithm, fees are properly calculated and observed during path finding. If a path has sufficient capacity _before_ fees are applied, but afterwards the finalized route would exceed the capacity of a single link, the route is marked as invalid. Currently a naive weighting algorithm is used which only factors in the time-lock delta at each hop, thereby optimizing for the lowest time lock. Fee calculation also isn’t finalized since we aren’t yet using milli-satoshi throughout the daemon. The final TODO item within the PR is to properly perform a multi-path search and rank the results based on a summation heuristic rather than just return the first (out of many) route found. On the server side, once nodes are initially connected to the daemon, our routing table will be synced with the peer’s using a naive “just send everything scheme” to hold us over until I spec out some a efficient graph reconciliation protocol. Additionally, the routing table is now pruned by the channel router itself once new blocks arrive rather than depending on peers to tell us when a channel flaps or is closed. Finally, the validation of peer announcements aren’t yet fully implemented as they’ll be implemented within the pending discovery package that was blocking on the completion of this package. Most off the routing message processing will be moved out of this package and into the discovery package where full validation will be carried out.
2016-12-27 08:20:26 +03:00
// We'll take the total timelock of the preceding hop as
// the outgoing timelock or this hop. Then we'll
// increment the total timelock incurred by this hop.
outgoingTimeLock = totalTimeLock
totalTimeLock += uint32(pathEdges[i+1].TimeLockDelta)
}
routing: rewrite package to conform to BOLT07 and factor in fees+timelocks This commit overhauls the routing package significantly to simplify the code, conform to the rest of the coding style within the package, and observe the new authenticated gossiping scheme outlined in BOLT07. As a major step towards a more realistic path finding algorithm, fees are properly calculated and observed during path finding. If a path has sufficient capacity _before_ fees are applied, but afterwards the finalized route would exceed the capacity of a single link, the route is marked as invalid. Currently a naive weighting algorithm is used which only factors in the time-lock delta at each hop, thereby optimizing for the lowest time lock. Fee calculation also isn’t finalized since we aren’t yet using milli-satoshi throughout the daemon. The final TODO item within the PR is to properly perform a multi-path search and rank the results based on a summation heuristic rather than just return the first (out of many) route found. On the server side, once nodes are initially connected to the daemon, our routing table will be synced with the peer’s using a naive “just send everything scheme” to hold us over until I spec out some a efficient graph reconciliation protocol. Additionally, the routing table is now pruned by the channel router itself once new blocks arrive rather than depending on peers to tell us when a channel flaps or is closed. Finally, the validation of peer announcements aren’t yet fully implemented as they’ll be implemented within the pending discovery package that was blocking on the completion of this package. Most off the routing message processing will be moved out of this package and into the discovery package where full validation will be carried out.
2016-12-27 08:20:26 +03:00
2018-10-24 03:16:39 +03:00
// Since we're traversing the path backwards atm, we prepend
// each new hop such that, the final slice of hops will be in
// the forwards order.
currentHop := &route.Hop{
PubKeyBytes: edge.Node.PubKeyBytes,
ChannelID: edge.ChannelID,
AmtToForward: amtToForward,
OutgoingTimeLock: outgoingTimeLock,
LegacyPayload: !tlvPayload,
CustomRecords: customRecords,
MPP: mpp,
}
hops = append([]*route.Hop{currentHop}, hops...)
2018-10-24 03:16:39 +03:00
// Finally, we update the amount that needs to flow into the
// *next* hop, which is the amount this hop needs to forward,
// accounting for the fee that it takes.
nextIncomingAmount = amtToForward + fee
routing: rewrite package to conform to BOLT07 and factor in fees+timelocks This commit overhauls the routing package significantly to simplify the code, conform to the rest of the coding style within the package, and observe the new authenticated gossiping scheme outlined in BOLT07. As a major step towards a more realistic path finding algorithm, fees are properly calculated and observed during path finding. If a path has sufficient capacity _before_ fees are applied, but afterwards the finalized route would exceed the capacity of a single link, the route is marked as invalid. Currently a naive weighting algorithm is used which only factors in the time-lock delta at each hop, thereby optimizing for the lowest time lock. Fee calculation also isn’t finalized since we aren’t yet using milli-satoshi throughout the daemon. The final TODO item within the PR is to properly perform a multi-path search and rank the results based on a summation heuristic rather than just return the first (out of many) route found. On the server side, once nodes are initially connected to the daemon, our routing table will be synced with the peer’s using a naive “just send everything scheme” to hold us over until I spec out some a efficient graph reconciliation protocol. Additionally, the routing table is now pruned by the channel router itself once new blocks arrive rather than depending on peers to tell us when a channel flaps or is closed. Finally, the validation of peer announcements aren’t yet fully implemented as they’ll be implemented within the pending discovery package that was blocking on the completion of this package. Most off the routing message processing will be moved out of this package and into the discovery package where full validation will be carried out.
2016-12-27 08:20:26 +03:00
}
// With the base routing data expressed as hops, build the full route
newRoute, err := route.NewRouteFromHops(
nextIncomingAmount, totalTimeLock, route.Vertex(sourceVertex),
hops,
2018-10-24 03:16:39 +03:00
)
if err != nil {
return nil, err
}
return newRoute, nil
}
routing: rewrite package to conform to BOLT07 and factor in fees+timelocks This commit overhauls the routing package significantly to simplify the code, conform to the rest of the coding style within the package, and observe the new authenticated gossiping scheme outlined in BOLT07. As a major step towards a more realistic path finding algorithm, fees are properly calculated and observed during path finding. If a path has sufficient capacity _before_ fees are applied, but afterwards the finalized route would exceed the capacity of a single link, the route is marked as invalid. Currently a naive weighting algorithm is used which only factors in the time-lock delta at each hop, thereby optimizing for the lowest time lock. Fee calculation also isn’t finalized since we aren’t yet using milli-satoshi throughout the daemon. The final TODO item within the PR is to properly perform a multi-path search and rank the results based on a summation heuristic rather than just return the first (out of many) route found. On the server side, once nodes are initially connected to the daemon, our routing table will be synced with the peer’s using a naive “just send everything scheme” to hold us over until I spec out some a efficient graph reconciliation protocol. Additionally, the routing table is now pruned by the channel router itself once new blocks arrive rather than depending on peers to tell us when a channel flaps or is closed. Finally, the validation of peer announcements aren’t yet fully implemented as they’ll be implemented within the pending discovery package that was blocking on the completion of this package. Most off the routing message processing will be moved out of this package and into the discovery package where full validation will be carried out.
2016-12-27 08:20:26 +03:00
// edgeWeight computes the weight of an edge. This value is used when searching
// for the shortest path within the channel graph between two nodes. Weight is
2018-07-31 10:19:49 +03:00
// is the fee itself plus a time lock penalty added to it. This benefits
// channels with shorter time lock deltas and shorter (hops) routes in general.
// RiskFactor controls the influence of time lock on route selection. This is
// currently a fixed value, but might be configurable in the future.
2018-06-04 23:10:05 +03:00
func edgeWeight(lockedAmt lnwire.MilliSatoshi, fee lnwire.MilliSatoshi,
timeLockDelta uint16) int64 {
// timeLockPenalty is the penalty for the time lock delta of this channel.
// It is controlled by RiskFactorBillionths and scales proportional
// to the amount that will pass through channel. Rationale is that it if
// a twice as large amount gets locked up, it is twice as bad.
2018-06-04 23:10:05 +03:00
timeLockPenalty := int64(lockedAmt) * int64(timeLockDelta) *
RiskFactorBillionths / 1000000000
2018-06-04 23:10:05 +03:00
return int64(fee) + timeLockPenalty
routing: rewrite package to conform to BOLT07 and factor in fees+timelocks This commit overhauls the routing package significantly to simplify the code, conform to the rest of the coding style within the package, and observe the new authenticated gossiping scheme outlined in BOLT07. As a major step towards a more realistic path finding algorithm, fees are properly calculated and observed during path finding. If a path has sufficient capacity _before_ fees are applied, but afterwards the finalized route would exceed the capacity of a single link, the route is marked as invalid. Currently a naive weighting algorithm is used which only factors in the time-lock delta at each hop, thereby optimizing for the lowest time lock. Fee calculation also isn’t finalized since we aren’t yet using milli-satoshi throughout the daemon. The final TODO item within the PR is to properly perform a multi-path search and rank the results based on a summation heuristic rather than just return the first (out of many) route found. On the server side, once nodes are initially connected to the daemon, our routing table will be synced with the peer’s using a naive “just send everything scheme” to hold us over until I spec out some a efficient graph reconciliation protocol. Additionally, the routing table is now pruned by the channel router itself once new blocks arrive rather than depending on peers to tell us when a channel flaps or is closed. Finally, the validation of peer announcements aren’t yet fully implemented as they’ll be implemented within the pending discovery package that was blocking on the completion of this package. Most off the routing message processing will be moved out of this package and into the discovery package where full validation will be carried out.
2016-12-27 08:20:26 +03:00
}
// graphParams wraps the set of graph parameters passed to findPath.
type graphParams struct {
// graph is the ChannelGraph to be used during path finding.
graph routingGraph
// additionalEdges is an optional set of edges that should be
// considered during path finding, that is not already found in the
// channel graph.
additionalEdges map[route.Vertex][]*channeldb.ChannelEdgePolicy
// bandwidthHints is an optional map from channels to bandwidths that
// can be populated if the caller has a better estimate of the current
// channel bandwidth than what is found in the graph. If set, it will
// override the capacities and disabled flags found in the graph for
// local channels when doing path finding. In particular, it should be
// set to the current available sending bandwidth for active local
// channels, and 0 for inactive channels.
bandwidthHints map[uint64]lnwire.MilliSatoshi
}
// RestrictParams wraps the set of restrictions passed to findPath that the
// found path must adhere to.
type RestrictParams struct {
// ProbabilitySource is a callback that is expected to return the
// success probability of traversing the channel from the node.
ProbabilitySource func(route.Vertex, route.Vertex,
lnwire.MilliSatoshi) float64
// FeeLimit is a maximum fee amount allowed to be used on the path from
// the source to the target.
FeeLimit lnwire.MilliSatoshi
// OutgoingChannelID is the channel that needs to be taken to the first
// hop. If nil, any channel may be used.
OutgoingChannelID *uint64
2019-11-18 13:54:15 +03:00
// LastHop is the pubkey of the last node before the final destination
// is reached. If nil, any node may be used.
LastHop *route.Vertex
// CltvLimit is the maximum time lock of the route excluding the final
// ctlv. After path finding is complete, the caller needs to increase
// all cltv expiry heights with the required final cltv delta.
CltvLimit uint32
// DestCustomRecords contains the custom records to drop off at the
// final hop, if any.
DestCustomRecords record.CustomSet
// DestFeatures is a feature vector describing what the final hop
// supports. If none are provided, pathfinding will try to inspect any
// features on the node announcement instead.
DestFeatures *lnwire.FeatureVector
// PaymentAddr is a random 32-byte value generated by the receiver to
// mitigate probing vectors and payment sniping attacks on overpaid
// invoices.
PaymentAddr *[32]byte
}
// PathFindingConfig defines global parameters that control the trade-off in
// path finding between fees and probabiity.
type PathFindingConfig struct {
// PaymentAttemptPenalty is the virtual cost in path finding weight
// units of executing a payment attempt that fails. It is used to trade
// off potentially better routes against their probability of
// succeeding.
PaymentAttemptPenalty lnwire.MilliSatoshi
// MinProbability defines the minimum success probability of the
// returned route.
MinProbability float64
}
2019-11-21 13:59:17 +03:00
// getMaxOutgoingAmt returns the maximum available balance in any of the
// channels of the given node.
func getMaxOutgoingAmt(node route.Vertex, outgoingChan *uint64,
2020-01-27 14:33:53 +03:00
bandwidthHints map[uint64]lnwire.MilliSatoshi,
g routingGraph) (lnwire.MilliSatoshi, error) {
2019-11-21 13:59:17 +03:00
var max lnwire.MilliSatoshi
2020-01-27 14:33:53 +03:00
cb := func(edgeInfo *channeldb.ChannelEdgeInfo, outEdge,
2019-11-21 13:59:17 +03:00
_ *channeldb.ChannelEdgePolicy) error {
if outEdge == nil {
return nil
}
chanID := outEdge.ChannelID
// Enforce outgoing channel restriction.
if outgoingChan != nil && chanID != *outgoingChan {
return nil
}
2020-01-27 14:33:53 +03:00
bandwidth, ok := bandwidthHints[chanID]
2019-11-21 13:59:17 +03:00
// If the bandwidth is not available for whatever reason, don't
// fail the pathfinding early.
if !ok {
max = lnwire.MaxMilliSatoshi
return nil
}
if bandwidth > max {
max = bandwidth
}
return nil
}
// Iterate over all channels of the to node.
2020-01-27 14:33:53 +03:00
err := g.forEachNodeChannel(node, cb)
2019-11-21 13:59:17 +03:00
if err != nil {
return 0, err
}
return max, err
}
2020-01-27 14:33:53 +03:00
// findPath attempts to find a path from the source node within the ChannelGraph
// to the target node that's capable of supporting a payment of `amt` value. The
// current approach implemented is modified version of Dijkstra's algorithm to
// find a single shortest path between the source node and the destination. The
// distance metric used for edges is related to the time-lock+fee costs along a
// particular edge. If a path is found, this function returns a slice of
// ChannelHop structs which encoded the chosen path from the target to the
// source. The search is performed backwards from destination node back to
// source. This is to properly accumulate fees that need to be paid along the
// path and accurately check the amount to forward at every node against the
// available bandwidth.
func findPath(g *graphParams, r *RestrictParams, cfg *PathFindingConfig,
source, target route.Vertex, amt lnwire.MilliSatoshi,
finalHtlcExpiry int32) ([]*channeldb.ChannelEdgePolicy, error) {
2020-01-27 14:33:53 +03:00
2019-09-06 09:56:59 +03:00
// Pathfinding can be a significant portion of the total payment
// latency, especially on low-powered devices. Log several metrics to
// aid in the analysis performance problems in this area.
start := time.Now()
nodesVisited := 0
edgesExpanded := 0
defer func() {
timeElapsed := time.Since(start)
log.Debugf("Pathfinding perf metrics: nodes=%v, edges=%v, "+
"time=%v", nodesVisited, edgesExpanded, timeElapsed)
}()
// If no destination features are provided, we will load what features
// we have for the target node from our graph.
features := r.DestFeatures
if features == nil {
2020-01-27 14:33:53 +03:00
var err error
features, err = g.graph.fetchNodeFeatures(target)
2020-01-27 14:33:53 +03:00
if err != nil {
return nil, err
}
routing: rewrite package to conform to BOLT07 and factor in fees+timelocks This commit overhauls the routing package significantly to simplify the code, conform to the rest of the coding style within the package, and observe the new authenticated gossiping scheme outlined in BOLT07. As a major step towards a more realistic path finding algorithm, fees are properly calculated and observed during path finding. If a path has sufficient capacity _before_ fees are applied, but afterwards the finalized route would exceed the capacity of a single link, the route is marked as invalid. Currently a naive weighting algorithm is used which only factors in the time-lock delta at each hop, thereby optimizing for the lowest time lock. Fee calculation also isn’t finalized since we aren’t yet using milli-satoshi throughout the daemon. The final TODO item within the PR is to properly perform a multi-path search and rank the results based on a summation heuristic rather than just return the first (out of many) route found. On the server side, once nodes are initially connected to the daemon, our routing table will be synced with the peer’s using a naive “just send everything scheme” to hold us over until I spec out some a efficient graph reconciliation protocol. Additionally, the routing table is now pruned by the channel router itself once new blocks arrive rather than depending on peers to tell us when a channel flaps or is closed. Finally, the validation of peer announcements aren’t yet fully implemented as they’ll be implemented within the pending discovery package that was blocking on the completion of this package. Most off the routing message processing will be moved out of this package and into the discovery package where full validation will be carried out.
2016-12-27 08:20:26 +03:00
}
// Ensure that the destination's features don't include unknown
// required features.
2020-01-27 14:33:53 +03:00
err := feature.ValidateRequired(features)
if err != nil {
return nil, err
}
// Ensure that all transitive dependencies are set.
err = feature.ValidateDeps(features)
if err != nil {
return nil, err
}
// Now that we know the feature vector is well formed, we'll proceed in
// checking that it supports the features we need, given our
// restrictions on the final hop.
// If the caller needs to send custom records, check that our
// destination feature vector supports TLV.
if len(r.DestCustomRecords) > 0 &&
!features.HasFeature(lnwire.TLVOnionPayloadOptional) {
return nil, errNoTlvPayload
}
// If the caller has a payment address to attach, check that our
// destination feature vector supports them.
if r.PaymentAddr != nil &&
!features.HasFeature(lnwire.PaymentAddrOptional) {
return nil, errNoPaymentAddr
}
2019-11-21 13:59:17 +03:00
// If we are routing from ourselves, check that we have enough local
// balance available.
self := g.graph.sourceNode()
2020-01-27 14:33:53 +03:00
2019-11-21 13:59:17 +03:00
if source == self {
2020-01-27 14:33:53 +03:00
max, err := getMaxOutgoingAmt(
self, r.OutgoingChannelID, g.bandwidthHints, g.graph,
2020-01-27 14:33:53 +03:00
)
2019-11-21 13:59:17 +03:00
if err != nil {
return nil, err
}
if max < amt {
return nil, errNoPathFound
2019-11-21 13:59:17 +03:00
}
}
// First we'll initialize an empty heap which'll help us to quickly
// locate the next edge we should visit next during our graph
// traversal.
nodeHeap := newDistanceHeap(estimatedNodeCount)
// Holds the current best distance for a given node.
distance := make(map[route.Vertex]*nodeWithDist, estimatedNodeCount)
additionalEdgesWithSrc := make(map[route.Vertex][]*edgePolicyWithSource)
for vertex, outgoingEdgePolicies := range g.additionalEdges {
// Build reverse lookup to find incoming edges. Needed because
// search is taken place from target to source.
2018-06-04 23:10:05 +03:00
for _, outgoingEdgePolicy := range outgoingEdgePolicies {
toVertex := outgoingEdgePolicy.Node.PubKeyBytes
incomingEdgePolicy := &edgePolicyWithSource{
sourceNode: vertex,
2018-06-04 23:10:05 +03:00
edge: outgoingEdgePolicy,
}
additionalEdgesWithSrc[toVertex] =
append(additionalEdgesWithSrc[toVertex],
incomingEdgePolicy)
}
}
// Build a preliminary destination hop structure to obtain the payload
// size.
var mpp *record.MPP
if r.PaymentAddr != nil {
mpp = record.NewMPP(amt, *r.PaymentAddr)
}
finalHop := route.Hop{
AmtToForward: amt,
OutgoingTimeLock: uint32(finalHtlcExpiry),
CustomRecords: r.DestCustomRecords,
LegacyPayload: !features.HasFeature(
lnwire.TLVOnionPayloadOptional,
),
MPP: mpp,
}
// We can't always assume that the end destination is publicly
// advertised to the network so we'll manually include the target node.
2019-11-18 12:19:20 +03:00
// The target node charges no fee. Distance is set to 0, because this is
// the starting point of the graph traversal. We are searching backwards
// to get the fees first time right and correctly match channel
// bandwidth.
//
// Don't record the initial partial path in the distance map and reserve
// that key for the source key in the case we route to ourselves.
partialPath := &nodeWithDist{
2018-06-04 23:10:05 +03:00
dist: 0,
weight: 0,
node: target,
2018-06-04 23:10:05 +03:00
amountToReceive: amt,
incomingCltv: finalHtlcExpiry,
probability: 1,
routingInfoSize: finalHop.PayloadSize(0),
}
// Calculate the absolute cltv limit. Use uint64 to prevent an overflow
// if the cltv limit is MaxUint32.
absoluteCltvLimit := uint64(r.CltvLimit) + uint64(finalHtlcExpiry)
// processEdge is a helper closure that will be used to make sure edges
// satisfy our specific requirements.
processEdge := func(fromVertex route.Vertex,
fromFeatures *lnwire.FeatureVector,
edge *channeldb.ChannelEdgePolicy, toNodeDist *nodeWithDist) {
2019-09-06 09:56:59 +03:00
edgesExpanded++
2020-02-27 16:22:00 +03:00
// Calculate amount that the candidate node would have to send
// out.
amountToSend := toNodeDist.amountToReceive
// Request the success probability for this edge.
edgeProbability := r.ProbabilitySource(
fromVertex, toNodeDist.node, amountToSend,
)
log.Trace(newLogClosure(func() string {
return fmt.Sprintf("path finding probability: fromnode=%v,"+
2020-03-17 15:22:32 +03:00
" tonode=%v, amt=%v, probability=%v",
fromVertex, toNodeDist.node, amountToSend,
edgeProbability)
}))
2018-06-04 23:10:05 +03:00
// If the probability is zero, there is no point in trying.
if edgeProbability == 0 {
return
}
2018-06-04 23:10:05 +03:00
// Compute fee that fromVertex is charging. It is based on the
2018-06-04 23:10:05 +03:00
// amount that needs to be sent to the next node in the route.
//
2018-09-06 11:48:46 +03:00
// Source node has no predecessor to pay a fee. Therefore set
// fee to zero, because it should not be included in the fee
// limit check and edge weight.
2018-06-04 23:10:05 +03:00
//
// Also determine the time lock delta that will be added to the
// route if fromVertex is selected. If fromVertex is the source
// node, no additional timelock is required.
2018-06-04 23:10:05 +03:00
var fee lnwire.MilliSatoshi
var timeLockDelta uint16
if fromVertex != source {
2019-01-24 23:14:05 +03:00
fee = edge.ComputeFee(amountToSend)
2018-06-04 23:10:05 +03:00
timeLockDelta = edge.TimeLockDelta
}
incomingCltv := toNodeDist.incomingCltv + int32(timeLockDelta)
// Check that we are within our CLTV limit.
if uint64(incomingCltv) > absoluteCltvLimit {
return
}
// amountToReceive is the amount that the node that is added to
// the distance map needs to receive from a (to be found)
// previous node in the route. That previous node will need to
// pay the amount that this node forwards plus the fee it
// charges.
2018-06-04 23:10:05 +03:00
amountToReceive := amountToSend + fee
// Check if accumulated fees would exceed fee limit when this
// node would be added to the path.
2018-06-04 23:10:05 +03:00
totalFee := amountToReceive - amt
if totalFee > r.FeeLimit {
2018-06-04 23:10:05 +03:00
return
}
// Calculate total probability of successfully reaching target
// by multiplying the probabilities. Both this edge and the rest
// of the route must succeed.
probability := toNodeDist.probability * edgeProbability
// If the probability is below the specified lower bound, we can
// abandon this direction. Adding further nodes can only lower
// the probability more.
if probability < cfg.MinProbability {
return
}
// By adding fromVertex in the route, there will be an extra
2018-06-04 23:10:05 +03:00
// weight composed of the fee that this node will charge and
// the amount that will be locked for timeLockDelta blocks in
// the HTLC that is handed out to fromVertex.
2018-06-04 23:10:05 +03:00
weight := edgeWeight(amountToReceive, fee, timeLockDelta)
// Compute the tentative weight to this new channel/edge
// which is the weight from our toNode to the target node
// plus the weight of this edge.
tempWeight := toNodeDist.weight + weight
// Add an extra factor to the weight to take into account the
// probability.
tempDist := getProbabilityBasedDist(
tempWeight, probability,
int64(cfg.PaymentAttemptPenalty),
)
// If there is already a best route stored, compare this
// candidate route with the best route so far.
current, ok := distance[fromVertex]
if ok {
// If this route is worse than what we already found,
// skip this route.
if tempDist > current.dist {
return
}
// If the route is equally good and the probability
// isn't better, skip this route. It is important to
// also return if both cost and probability are equal,
// because otherwise the algorithm could run into an
// endless loop.
probNotBetter := probability <= current.probability
if tempDist == current.dist && probNotBetter {
return
}
2018-06-04 23:10:05 +03:00
}
// Every edge should have a positive time lock delta. If we
// encounter a zero delta, log a warning line.
2018-06-04 23:10:05 +03:00
if edge.TimeLockDelta == 0 {
log.Warnf("Channel %v has zero cltv delta",
edge.ChannelID)
}
2018-06-04 23:10:05 +03:00
// Calculate the total routing info size if this hop were to be
// included. If we are coming from the source hop, the payload
// size is zero, because the original htlc isn't in the onion
// blob.
var payloadSize uint64
if fromVertex != source {
supportsTlv := fromFeatures.HasFeature(
lnwire.TLVOnionPayloadOptional,
)
hop := route.Hop{
AmtToForward: amountToSend,
OutgoingTimeLock: uint32(
toNodeDist.incomingCltv,
),
LegacyPayload: !supportsTlv,
}
payloadSize = hop.PayloadSize(edge.ChannelID)
}
routingInfoSize := toNodeDist.routingInfoSize + payloadSize
// Skip paths that would exceed the maximum routing info size.
if routingInfoSize > sphinx.MaxPayloadSize {
return
}
2018-06-04 23:10:05 +03:00
// All conditions are met and this new tentative distance is
// better than the current best known distance to this node.
// The new better distance is recorded, and also our "next hop"
// map is populated with this edge.
withDist := &nodeWithDist{
2018-06-04 23:10:05 +03:00
dist: tempDist,
weight: tempWeight,
node: fromVertex,
2018-06-04 23:10:05 +03:00
amountToReceive: amountToReceive,
incomingCltv: incomingCltv,
probability: probability,
nextHop: edge,
routingInfoSize: routingInfoSize,
2018-06-04 23:10:05 +03:00
}
distance[fromVertex] = withDist
2018-06-04 23:10:05 +03:00
// Either push withDist onto the heap if the node
// represented by fromVertex is not already on the heap OR adjust
// its position within the heap via heap.Fix.
nodeHeap.PushOrFix(withDist)
}
// TODO(roasbeef): also add path caching
// * similar to route caching, but doesn't factor in the amount
// Cache features because we visit nodes multiple times.
featureCache := make(map[route.Vertex]*lnwire.FeatureVector)
// getGraphFeatures returns (cached) node features from the graph.
getGraphFeatures := func(node route.Vertex) (*lnwire.FeatureVector,
error) {
// Check cache for features of the fromNode.
fromFeatures, ok := featureCache[node]
2020-01-27 14:33:53 +03:00
if ok {
return fromFeatures, nil
}
2020-01-27 14:33:53 +03:00
// Fetch node features fresh from the graph.
fromFeatures, err := g.graph.fetchNodeFeatures(node)
2020-01-27 14:33:53 +03:00
if err != nil {
return nil, err
}
2020-01-27 14:33:53 +03:00
// Don't route through nodes that contain unknown required
// features and mark as nil in the cache.
err = feature.ValidateRequired(fromFeatures)
if err != nil {
featureCache[node] = nil
return nil, nil
}
2020-01-27 14:33:53 +03:00
// Don't route through nodes that don't properly set all
// transitive feature dependencies and mark as nil in the cache.
err = feature.ValidateDeps(fromFeatures)
if err != nil {
featureCache[node] = nil
return nil, nil
}
// Update cache.
featureCache[node] = fromFeatures
return fromFeatures, nil
}
2019-11-18 12:19:20 +03:00
routeToSelf := source == target
for {
2019-09-06 09:56:59 +03:00
nodesVisited++
pivot := partialPath.node
routing: rewrite package to conform to BOLT07 and factor in fees+timelocks This commit overhauls the routing package significantly to simplify the code, conform to the rest of the coding style within the package, and observe the new authenticated gossiping scheme outlined in BOLT07. As a major step towards a more realistic path finding algorithm, fees are properly calculated and observed during path finding. If a path has sufficient capacity _before_ fees are applied, but afterwards the finalized route would exceed the capacity of a single link, the route is marked as invalid. Currently a naive weighting algorithm is used which only factors in the time-lock delta at each hop, thereby optimizing for the lowest time lock. Fee calculation also isn’t finalized since we aren’t yet using milli-satoshi throughout the daemon. The final TODO item within the PR is to properly perform a multi-path search and rank the results based on a summation heuristic rather than just return the first (out of many) route found. On the server side, once nodes are initially connected to the daemon, our routing table will be synced with the peer’s using a naive “just send everything scheme” to hold us over until I spec out some a efficient graph reconciliation protocol. Additionally, the routing table is now pruned by the channel router itself once new blocks arrive rather than depending on peers to tell us when a channel flaps or is closed. Finally, the validation of peer announcements aren’t yet fully implemented as they’ll be implemented within the pending discovery package that was blocking on the completion of this package. Most off the routing message processing will be moved out of this package and into the discovery package where full validation will be carried out.
2016-12-27 08:20:26 +03:00
// Create unified policies for all incoming connections.
u := newUnifiedPolicies(self, pivot, r.OutgoingChannelID)
2018-06-04 23:10:05 +03:00
err := u.addGraphPolicies(g.graph)
if err != nil {
return nil, err
}
for _, reverseEdge := range additionalEdgesWithSrc[pivot] {
u.addPolicy(reverseEdge.sourceNode, reverseEdge.edge, 0)
}
amtToSend := partialPath.amountToReceive
// Expand all connections using the optimal policy for each
// connection.
for fromNode, unifiedPolicy := range u.policies {
2019-11-18 12:19:20 +03:00
// The target node is not recorded in the distance map.
// Therefore we need to have this check to prevent
// creating a cycle. Only when we intend to route to
// self, we allow this cycle to form. In that case we'll
// also break out of the search loop below.
if !routeToSelf && fromNode == target {
continue
}
2019-11-18 13:54:15 +03:00
// Apply last hop restriction if set.
if r.LastHop != nil &&
pivot == target && fromNode != *r.LastHop {
continue
}
policy := unifiedPolicy.getPolicy(
amtToSend, g.bandwidthHints,
)
if policy == nil {
continue
2018-06-04 23:10:05 +03:00
}
// Get feature vector for fromNode.
fromFeatures, err := getGraphFeatures(fromNode)
if err != nil {
return nil, err
}
// If there are no valid features, skip this node.
if fromFeatures == nil {
continue
}
// Check if this candidate node is better than what we
// already have.
processEdge(fromNode, fromFeatures, policy, partialPath)
}
if nodeHeap.Len() == 0 {
break
}
// Fetch the node within the smallest distance from our source
// from the heap.
partialPath = heap.Pop(&nodeHeap).(*nodeWithDist)
// If we've reached our source (or we don't have any incoming
// edges), then we're done here and can exit the graph
// traversal early.
if partialPath.node == source {
break
}
routing: rewrite package to conform to BOLT07 and factor in fees+timelocks This commit overhauls the routing package significantly to simplify the code, conform to the rest of the coding style within the package, and observe the new authenticated gossiping scheme outlined in BOLT07. As a major step towards a more realistic path finding algorithm, fees are properly calculated and observed during path finding. If a path has sufficient capacity _before_ fees are applied, but afterwards the finalized route would exceed the capacity of a single link, the route is marked as invalid. Currently a naive weighting algorithm is used which only factors in the time-lock delta at each hop, thereby optimizing for the lowest time lock. Fee calculation also isn’t finalized since we aren’t yet using milli-satoshi throughout the daemon. The final TODO item within the PR is to properly perform a multi-path search and rank the results based on a summation heuristic rather than just return the first (out of many) route found. On the server side, once nodes are initially connected to the daemon, our routing table will be synced with the peer’s using a naive “just send everything scheme” to hold us over until I spec out some a efficient graph reconciliation protocol. Additionally, the routing table is now pruned by the channel router itself once new blocks arrive rather than depending on peers to tell us when a channel flaps or is closed. Finally, the validation of peer announcements aren’t yet fully implemented as they’ll be implemented within the pending discovery package that was blocking on the completion of this package. Most off the routing message processing will be moved out of this package and into the discovery package where full validation will be carried out.
2016-12-27 08:20:26 +03:00
}
// Use the distance map to unravel the forward path from source to
// target.
var pathEdges []*channeldb.ChannelEdgePolicy
currentNode := source
for {
2018-06-04 23:10:05 +03:00
// Determine the next hop forward using the next map.
currentNodeWithDist, ok := distance[currentNode]
if !ok {
// If the node doesnt have a next hop it means we didn't find a path.
return nil, errNoPathFound
}
2018-06-04 23:10:05 +03:00
// Add the next hop to the list of path edges.
pathEdges = append(pathEdges, currentNodeWithDist.nextHop)
2018-06-04 23:10:05 +03:00
// Advance current node.
currentNode = currentNodeWithDist.nextHop.Node.PubKeyBytes
2019-11-18 12:19:20 +03:00
// Check stop condition at the end of this loop. This prevents
// breaking out too soon for self-payments that have target set
// to source.
if currentNode == target {
break
}
}
// For the final hop, we'll set the node features to those determined
// above. These are either taken from the destination features, e.g.
// virtual or invoice features, or loaded as a fallback from the graph.
// The transitive dependencies were already validated above, so no need
// to do so now.
//
// NOTE: This may overwrite features loaded from the graph if
// destination features were provided. This is fine though, since our
// route construction does not care where the features are actually
// taken from. In the future we may wish to do route construction within
// findPath, and avoid using ChannelEdgePolicy altogether.
pathEdges[len(pathEdges)-1].Node.Features = features
2019-12-20 12:28:48 +03:00
log.Debugf("Found route: probability=%v, hops=%v, fee=%v",
distance[source].probability, len(pathEdges),
distance[source].amountToReceive-amt)
return pathEdges, nil
}
// getProbabilityBasedDist converts a weight into a distance that takes into
// account the success probability and the (virtual) cost of a failed payment
// attempt.
//
// Derivation:
//
// Suppose there are two routes A and B with fees Fa and Fb and success
// probabilities Pa and Pb.
//
// Is the expected cost of trying route A first and then B lower than trying the
// other way around?
//
// The expected cost of A-then-B is: Pa*Fa + (1-Pa)*Pb*(c+Fb)
//
// The expected cost of B-then-A is: Pb*Fb + (1-Pb)*Pa*(c+Fa)
//
// In these equations, the term representing the case where both A and B fail is
// left out because its value would be the same in both cases.
//
// Pa*Fa + (1-Pa)*Pb*(c+Fb) < Pb*Fb + (1-Pb)*Pa*(c+Fa)
//
// Pa*Fa + Pb*c + Pb*Fb - Pa*Pb*c - Pa*Pb*Fb < Pb*Fb + Pa*c + Pa*Fa - Pa*Pb*c - Pa*Pb*Fa
//
// Removing terms that cancel out:
// Pb*c - Pa*Pb*Fb < Pa*c - Pa*Pb*Fa
//
// Divide by Pa*Pb:
// c/Pa - Fb < c/Pb - Fa
//
// Move terms around:
// Fa + c/Pa < Fb + c/Pb
//
// So the value of F + c/P can be used to compare routes.
func getProbabilityBasedDist(weight int64, probability float64, penalty int64) int64 {
// Clamp probability to prevent overflow.
const minProbability = 0.00001
if probability < minProbability {
return infinity
}
return weight + int64(float64(penalty)/probability)
}