[litep2p](https://github.com/altonen/litep2p) is a libp2p-compatible P2P
networking library. It supports all of the features of `rust-libp2p`
that are currently being utilized by Polkadot SDK.
Compared to `rust-libp2p`, `litep2p` has a quite different architecture
which is why the new `litep2p` network backend is only able to use a
little of the existing code in `sc-network`. The design has been mainly
influenced by how we'd wish to structure our networking-related code in
Polkadot SDK: independent higher-levels protocols directly communicating
with the network over links that support bidirectional backpressure. A
good example would be `NotificationHandle`/`RequestResponseHandle`
abstractions which allow, e.g., `SyncingEngine` to directly communicate
with peers to announce/request blocks.
I've tried running `polkadot --network-backend litep2p` with a few
different peer configurations and there is a noticeable reduction in
networking CPU usage. For high load (`--out-peers 200`), networking CPU
usage goes down from ~110% to ~30% (80 pp) and for normal load
(`--out-peers 40`), the usage goes down from ~55% to ~18% (37 pp).
These should not be taken as final numbers because:
a) there are still some low-hanging optimization fruits, such as
enabling [receive window
auto-tuning](https://github.com/libp2p/rust-yamux/pull/176), integrating
`Peerset` more closely with `litep2p` or improving memory usage of the
WebSocket transport
b) fixing bugs/instabilities that incorrectly cause `litep2p` to do less
work will increase the networking CPU usage
c) verification in a more diverse set of tests/conditions is needed
Nevertheless, these numbers should give an early estimate for CPU usage
of the new networking backend.
This PR consists of three separate changes:
* introduce a generic `PeerId` (wrapper around `Multihash`) so that we
don't have use `NetworkService::PeerId` in every part of the code that
uses a `PeerId`
* introduce `NetworkBackend` trait, implement it for the libp2p network
stack and make Polkadot SDK generic over `NetworkBackend`
* implement `NetworkBackend` for litep2p
The new library should be considered experimental which is why
`rust-libp2p` will remain as the default option for the time being. This
PR currently depends on the master branch of `litep2p` but I'll cut a
new release for the library once all review comments have been
addresses.
---------
Signed-off-by: Alexandru Vasile <alexandru.vasile@parity.io>
Co-authored-by: Dmitry Markin <dmitry@markin.tech>
Co-authored-by: Alexandru Vasile <60601340+lexnv@users.noreply.github.com>
Co-authored-by: Alexandru Vasile <alexandru.vasile@parity.io>
# Description
Trivial change that resolves
https://github.com/paritytech/polkadot-sdk/issues/2185.
Since there was a mix of `who` and `peer_id` argument names nearby I
changed them all to `peer_id`.
# Checklist
- [x] My PR includes a detailed description as outlined in the
"Description" section above
- [x] My PR follows the [labeling requirements](CONTRIBUTING.md#Process)
of this project (at minimum one label for `T`
required)
- [x] I have made corresponding changes to the documentation (if
applicable)
- [ ] I have added tests that prove my fix is effective or that my
feature works (if applicable)
---------
Co-authored-by: Bastian Köcher <git@kchr.de>
This commit introduces a new concept called `NotificationService` which
allows Polkadot protocols to communicate with the underlying
notification protocol implementation directly, without routing events
through `NetworkWorker`. This implies that each protocol has its own
service which it uses to communicate with remote peers and that each
`NotificationService` is unique with respect to the underlying
notification protocol, meaning `NotificationService` for the transaction
protocol can only be used to send and receive transaction-related
notifications.
The `NotificationService` concept introduces two additional benefits:
* allow protocols to start using custom handshakes
* allow protocols to accept/reject inbound peers
Previously the validation of inbound connections was solely the
responsibility of `ProtocolController`. This caused issues with light
peers and `SyncingEngine` as `ProtocolController` would accept more
peers than `SyncingEngine` could accept which caused peers to have
differing views of their own states. `SyncingEngine` would reject excess
peers but these rejections were not properly communicated to those peers
causing them to assume that they were accepted.
With `NotificationService`, the local handshake is not sent to remote
peer if peer is rejected which allows it to detect that it was rejected.
This commit also deprecates the use of `NetworkEventStream` for all
notification-related events and going forward only DHT events are
provided through `NetworkEventStream`. If protocols wish to follow each
other's events, they must introduce additional abtractions, as is done
for GRANDPA and transactions protocols by following the syncing protocol
through `SyncEventStream`.
Fixes https://github.com/paritytech/polkadot-sdk/issues/512
Fixes https://github.com/paritytech/polkadot-sdk/issues/514
Fixes https://github.com/paritytech/polkadot-sdk/issues/515
Fixes https://github.com/paritytech/polkadot-sdk/issues/554
Fixes https://github.com/paritytech/polkadot-sdk/issues/556
---
These changes are transferred from
https://github.com/paritytech/substrate/pull/14197 but there are no
functional changes compared to that PR
---------
Co-authored-by: Dmitry Markin <dmitry@markin.tech>
Co-authored-by: Alexandru Vasile <60601340+lexnv@users.noreply.github.com>
This PR moves syncing-related code from `sc-network-common` to
`sc-network-sync`.
Unfortunately, some parts are tightly integrated with networking, so
they were left in `sc-network-common` for now:
1. `SyncMode` in `common/src/sync.rs` (used in `NetworkConfiguration`).
2. `BlockAnnouncesHandshake`, `BlockRequest`, `BlockResponse`, etc. in
`common/src/sync/message.rs` (used in `src/protocol.rs` and
`src/protocol/message.rs`).
More substantial refactoring is needed to decouple syncing and
networking completely, including getting rid of the hardcoded sync
protocol.
## Release notes
Move syncing-related code from `sc-network-common` to `sc-network-sync`.
Delete `ChainSync` trait as it's never used (the only implementation is
accessed directly from `SyncingEngine` and exposes a lot of public
methods that are not part of the trait). Some new trait(s) for syncing
will likely be introduced as part of Sync 2.0 refactoring to represent
syncing strategies.
* update libp2p to 0.52.0
* proto name now must implement `AsRef<str>`
* update libp2p version everywhere
* ToSwarm, FromBehaviour, ToBehaviour
also LocalProtocolsChange and RemoteProtocolsChange
* new NetworkBehaviour invariants
* replace `Vec<u8>` with `StreamProtocol`
* rename ConnectionHandlerEvent::Custom to NotifyBehaviour
* remove DialError & ListenError invariants
also fix pending_events
* use connection_limits::Behaviour
See https://github.com/libp2p/rust-libp2p/pull/3885
* impl `void::Void` for `BehaviourOut`
also use `Behaviour::with_codec`
* KademliaHandler no longer public
* fix StreamProtocol construction
* update libp2p-identify to 0.2.0
* remove non-existing methods from PollParameters
rename ConnectionHandlerUpgrErr to StreamUpgradeError
* `P2p` now contains `PeerId`, not `Multihash`
* use multihash-codetable crate
* update Cargo.lock
* reformat text
* comment out tests for now
* remove `.into()` from P2p
* confirm observed addr manually
See https://github.com/libp2p/rust-libp2p/blob/master/protocols/identify/CHANGELOG.md#0430
* remove SwarmEvent::Banned
since we're not using `ban_peer_id`, this can be safely removed.
we may want to introduce `libp2p::allow_block_list` module in the future.
* fix imports
* replace `libp2p` with smaller deps in network-gossip
* bring back tests
* finish rewriting tests
* uncomment handler tests
* Revert "uncomment handler tests"
This reverts commit 720a06815887f4e10767c62b58864a7ec3a48e50.
* add a fixme
* update Cargo.lock
* remove extra From
* make void uninhabited
* fix discovery test
* use autonat protocols
confirming external addresses manually is unsafe in open networks
* fix SyncNotificationsClogged invariant
* only set server mode manually in tests
doubt that we need to set it on node since we're adding public addresses
* address @dmitry-markin comments
* remove autonat
* removed unused var
* fix EOL
* update smallvec and sha2
in attempt to compile polkadot
* bump k256
in attempt to build cumulus
---------
Co-authored-by: parity-processbot <>
* Move service tests to `client/network/tests`
These tests depend on `sc-network` and `sc-network-sync` so they should
live outside the crate.
* Move some configs from `sc-network-common` to `sc-network`
* Move `NetworkService` traits to `sc-network`
* Move request-responses to `sc-network`
* Remove more stuff
* Remove rest of configs from `sc-network-common` to `sc-network`
* Remove more stuff
* Fix warnings
* Update client/network/src/request_responses.rs
Co-authored-by: Dmitry Markin <dmitry@markin.tech>
* Fix cargo doc
---------
Co-authored-by: Dmitry Markin <dmitry@markin.tech>
* Move import queue out of `sc-network`
Add supplementary asynchronous API for the import queue which means
it can be run as an independent task and communicated with through
the `ImportQueueService`.
This commit removes removes block and justification imports from
`sc-network` and provides `ChainSync` with a handle to import queue so
it can import blocks and justifications. Polling of the import queue is
moved complete out of `sc-network` and `sc_consensus::Link` is
implemented for `ChainSyncInterfaceHandled` so the import queue
can still influence the syncing process.
* Move stuff to SyncingEngine
* Move `ChainSync` instanation to `SyncingEngine`
Some of the tests have to be rewritten
* Move peer hashmap to `SyncingEngine`
* Let `SyncingEngine` to implement `ChainSyncInterface`
* Introduce `SyncStatusProvider`
* Move `sync_peer_(connected|disconnected)` to `SyncingEngine`
* Implement `SyncEventStream`
Remove `SyncConnected`/`SyncDisconnected` events from
`NetworkEvenStream` and provide those events through
`ChainSyncInterface` instead.
Modify BEEFY/GRANDPA/transactions protocol and `NetworkGossip` to take
`SyncEventStream` object which they listen to for incoming sync peer
events.
* Introduce `ChainSyncInterface`
This interface provides a set of miscellaneous functions that other
subsystems can use to query, for example, the syncing status.
* Move event stream polling to `SyncingEngine`
Subscribe to `NetworkStreamEvent` and poll the incoming notifications
and substream events from `SyncingEngine`.
The code needs refactoring.
* Make `SyncingEngine` into an asynchronous runner
This commits removes the last hard dependency of syncing from
`sc-network` meaning the protocol now lives completely outside of
`sc-network`, ignoring the hardcoded peerset entry which will be
addressed in the future.
Code needs a lot of refactoring.
* Fix warnings
* Code refactoring
* Use `SyncingService` for BEEFY
* Use `SyncingService` for GRANDPA
* Remove call delegation from `NetworkService`
* Remove `ChainSyncService`
* Remove `ChainSync` service tests
They were written for the sole purpose of verifying that `NetworWorker`
continues to function while the calls are being dispatched to
`ChainSync`.
* Refactor code
* Refactor code
* Update client/finality-grandpa/src/communication/tests.rs
Co-authored-by: Anton <anton.kalyaev@gmail.com>
* Fix warnings
* Apply review comments
* Fix docs
* Fix test
* cargo-fmt
* Update client/network/sync/src/engine.rs
Co-authored-by: Anton <anton.kalyaev@gmail.com>
* Update client/network/sync/src/engine.rs
Co-authored-by: Anton <anton.kalyaev@gmail.com>
* Add missing docs
* Refactor code
---------
Co-authored-by: Anton <anton.kalyaev@gmail.com>
* Change copyright year to 2023 from 2022
* Fix incorrect update of copyright year
* Remove years from copy right header
* Fix remaining files
* Fix typo in a header and remove update-copyright.sh
* Replace deprecated libp2p feature specs with correct ones
* Bump tokio to 1.21.2
* Replace async-std libp2p primitives with tokio ones
* minor: rustfmt
* Fix TestNet to run initialization in the tokio context
* Convert telemetry test from async-std to tokio
* Convert notifications tests from async-std to tokio
* Convert chain sync tests from async-std to tokio
* Ditch async-std completely
* Make executor mandatory
* Bump tokio to 1.22.0
* minor: rustfmt
* Explicitly use tokio runtime in tests
* Move more tests to explicit tokio runtime
* Explicitly set multithreaded runtime in tokio test
* minor: rustfmt
* minor: fix comment
* Replace async-std with tokio in MMR tests
* Add ProtocolName custom type
* Use new ProtocolName in sc_network_common
* Use new ProtocolName in sc_network
* Use new ProtocolName for BEEFY and GRANDPA
* Use new ProtocolName for notifications
* Use new ProtocolName in sc_network (part 2)
* Use new ProtocolName in sc_network_gossip
* Use new ProtocolName in sc_offchain
* Remove unused imports
* Some more fixes
* Add tests
* Fix minor import issues
* Re-export ProtocolName in sc_network
* Revert "Re-export ProtocolName in sc_network"
This reverts commit 8d8ff71927e7750757f29c9bbd88dc0ba181d214.
* Re-export ProtocolName in sc_network
* Remove dependency on sc-network-common from beefy-gadget
* Move `MultiaddrWithPeerId` and related parsing functions into `sc-network-common`, remove dependency on `sc-network` from `sc-chain-spec`
* Remove dependency on `sc-network` from `sc-offchain`
* Remove dependency on `sc-network` from `sc-network-gossip`
* Extract `NetworkKVProvider` trait in `sc-authority-discovery` and remove unnecessary dependency
* Extract `NetworkSyncForkRequest` trait in `sc-finality-grandpa`
* Relax requirements on `SyncOracle` trait, remove extra native methods from `NetworkService` that are already provided by trait impls
* Move `NetworkSigner` trait from `sc-authority-discovery` into `sc-network-common` and de-duplicate methods on `NetworkService`
* Move `NetworkKVProvider` trait from `sc-authority-discovery` into `sc-network-common` and de-duplicate methods on `NetworkService`
* Minimize `sc-authority-discovery` dependency on `sc-network`
* Move `NetworkSyncForkRequest` trait from `sc-finality-grandpa` to `sc-network-common` and de-duplicate methods in `NetworkService`
* Extract `NetworkStatusProvider` trait and de-duplicate methods on `NetworkService`
* Extract `NetworkPeers` trait and de-duplicate methods on `NetworkService`
* Extract `NetworkEventStream` trait and de-duplicate methods on `NetworkService`
* Move more methods from `NetworkService` into `NetworkPeers` trait
* Move `NetworkStateInfo` trait into `sc-network-common`
* Extract `NetworkNotification` trait and de-duplicate methods on `NetworkService`
* Extract `NetworkRequest` trait and de-duplicate methods on `NetworkService`
* Remove `NetworkService::local_peer_id()`, it is already provided by `NetworkStateInfo` impl
* Extract `NetworkTransaction` trait and de-duplicate methods on `NetworkService`
* Extract `NetworkBlock` trait and de-duplicate methods on `NetworkService`
* Remove dependencies on `NetworkService` from most of the methods of `sc-service`
* Address simple review comments
* Pump the gossip engine while waiting for the BEEFY runtime pallet
This fixes a memory leak when the BEEFY gadget is turned on, but
the runtime doesn't actually use BEEFY.
* Implement `FusedFuture` for `GossipEngine`
* Fuse futures outside of loops
Remove unneeded dependencies and dev-dependencies.
Made self_destruct test not dependent on wasm bin size.
Updated code related to deprecated warning on tracing-subscriber `scope()`
( See https://github.com/tokio-rs/tracing/issues/1429 )
* Run cargo fmt on the whole code base
* Second run
* Add CI check
* Fix compilation
* More unnecessary braces
* Handle weights
* Use --all
* Use correct attributes...
* Fix UI tests
* AHHHHHHHHH
* 🤦
* Docs
* Fix compilation
* 🤷
* Please stop
* 🤦 x 2
* More
* make rustfmt.toml consistent with polkadot
Co-authored-by: André Silva <andrerfosilva@gmail.com>
* Allow fallback names for protocols
* Apply suggestions from code review
Co-authored-by: Roman Proskuryakov <humbug@deeptown.org>
* Fix some issues
* Fix compilation after merging master
Co-authored-by: Roman Proskuryakov <humbug@deeptown.org>
* Sync: Propagate block announcement data
This pr adds a feature to the sync protocol to propagate the data that
we received alongside a block announcement. This is done by adding a
cache that caches the last X block announcement data where X is set to
the number of `in_peers` (giving every peer the chance to send us a
different block). This will be required by parachains to ensure that
even peers who are not connected to a collator receive the data
alongside the block announcement to properly validate it and request the
block.
* Review comment
* Bring back the code and add new variant to ensure we don't insert block
announce data when something wasn't checked
* Also use out_peers
* network-gossip: add metric for number of local messages
* grandpa: fix GossipEngine missing metrics registry parameter
* network-gossip: increase known messages cache size
* network-gossip: fix tests
* grandpa: remove unnecessary clone
Co-authored-by: Max Inden <mail@max-inden.de>
* network-gossip: count registered and expired messages separately
* network-gossip: add comment on known messages cache size
* network-gossip: extend comment with cache size in memory
Co-authored-by: Max Inden <mail@max-inden.de>
* Rework priority groups
* Broken tests fix
* Fix warning causing CI to fail
* [Hack] Try restore backwards-compatibility
* Fix peerset bug
* Doc fixes and clean up
* Error on state mismatch
* Try debug CI
* CI debugging
* [CI debug] Can I please see this line
* Revert "[CI debug] Can I please see this line"
This reverts commit 4b7cf7c1511f579cd818b21d46bd11642dfac5cb.
* Revert "CI debugging"
This reverts commit 9011f1f564b860386dc7dd6ffa9fc34ea7107623.
* Fix error! which isn't actually an error
* Fix Ok() returned when actually Err()
* Tweaks and fixes
* Fix build
* Peerset bugfix
* [Debug] Try outbound GrandPa slots
* Another bugfix
* Revert "[Debug] Try outbound GrandPa slots"
This reverts commit d175b9208c088faad77d9f0ce36ff6f48bd92dd3.
* [Debug] Try outbound GrandPa slots
* Apply suggestions from code review
Co-authored-by: Max Inden <mail@max-inden.de>
* Use consts for hardcoded peersets
* Revert "Try debug CI"
This reverts commit 62c4ad5e79c03d561c714a008022ecac463a597e.
* Renames
* Line widths
* Add doc
Co-authored-by: Max Inden <mail@max-inden.de>
* Remove sc_network::NetworkService::register_notifications_protocol
* Missing calls to .into()
* Wrong crate name
* [WIP] Fix Grandpa tests
* One more passing
* One more. Two to go.
* This one was actually already passing 🎉
* Last one compiles
* Progress
* grandpa: fix voter_persists_its_votes test
* Restore other tests
* Try spawn future later
Co-authored-by: André Silva <andrerfosilva@gmail.com>
* Remove necessity to pass ConsensusEngineId when registering notifications protocol
* Line width
* Fix tests protocol name
* Other renames
* Doc update
* Change issue in TODO
* client/*: Treat protocol name as str and not [u8]
Notification protocol names are in practice always valid utf8 strings.
Instead of treating them as such in the type system, thus far they were
casted to a [u8] at creation time.
With this commit protocol names are instead treated as valid utf8
strings throughout the codebase and passed as `Cow<'static, str>`
instead of `Cow<'static, [u8]>`. Among other things this eliminates the
need for string casting when logging.
* client/network: Don't allocate when protocol name is borrowed
* client/network-gossip/bridge: Use bounded channel
Instead of returning an unbounded channel on
`GossipEngine::messages_for` return a bounded channel. For now the
channel length is determined by the amount of past messages cached in
the `ConsensusGossip`.
With a bounded channel, one can't just fire-and-forget style send into
it, but has to first check whether the channel is ready. Thus this
commit restructures `GossipEngine::poll` and introduces a
`ForwardingState` into `GossipEngine`.
* client/network-gossip/bridge: Add quickcheck for different size channels
* client/network-gossip/src/state_machine: Add unit test for on_incoming
Add two unit tests to ensure `on_incoming` is ingoring discarded
messages and reports and ignores messages from unknown peers.
* client/network-gossip/state_machine: Reduce indentation level
* client/network-gossip/bridge: Remove unnecessary into_iter
* client/network-gossip/state_machine: Report success after register check
* client/network-gossip/state_machine: Error not report on unregistered
`peers` contains all the peers we're connected to. If we receive a
message from a peer not in this list, that means there's an internal
problem within the local client. It's not the fault of the peer in
question.
With the above in mind instead of reducing the reputation of such peer,
log an error.
* client/network-gossip: Move sink IO outside of state_machine
`ConsensusGossip` is supposed to be a deterministic state machine.
`GossipEngine` wrapping `ConsensusGossip` should handle IO operations.
This commit moves the `message_sink` IO operations to `GossipEngine`.
More specifically on incoming messages a `GossipEngine` calls
`ConsensusGossip::on_incoming` to validate and register the messages.
`ConsensusGossip` returns the valid messages which are then forwarded by
`GossipEngine` to the upper layer via the `message_sinks`.
* client/network-gossip: Adjust and extend tests
* Update client/network-gossip/src/bridge.rs
Co-authored-by: Benjamin Kampmann <ben.kampmann@googlemail.com>
This PR refactors the metrics measuring and Prometheus exposing entity in sc-service into its own submodule and extends the parameters it exposes by:
- system load average (over one, five and 15min)
- the TCP connection state of the process (lsof), refs #5304
- number of tokio threads
- number of known forks
- counter for items in each unbounded queue (with internal unbounded channels)
- number of file descriptors opened by this process (*nix only at this point)
- number of system threads (*nix only at this point)
refs #4679
Co-authored-by: Max Inden <mail@max-inden.de>
Co-authored-by: Ashley <ashley.ruglys@gmail.com>
* Split the Roles bitfield in three
* Forgot to include some changes
* Fix cli test
* More test fixes
* Oh God, merging master broke other tests
* Didn't run the doctests
* Address review
* I'm trying to fix the build blindly because it's taking a good hour to compile on my machine
* Address some review
* Also update the peerset's API to make sense
* Fix peerset tests
* Fix browser node
* client: distinguish between local and network authority
Co-authored-by: André Silva <andre.beat@gmail.com>
* client/network-gossip/src/bridge: Finish when network even stream closes
Previously within `<GossipEngine as Future>::poll` one would poll the
`network_event_stream` ignoring all messages other than
`Poll::Ready(Some())`. Ignoring `Poll::Ready(None)` leads to a panic on
the next poll of the stream, gien that it is not fused.
By design `network_event_stream` does not close unless an unbounded send
into it fails, or the `NetworkWorker` gets shut down.
> The stream never ends (unless the `NetworkWorker` gets shut down).
> (client/network/src/service.rs)
An `unbounded_send` to fail on an unbounded channel is unlikely. The
`NetworkWorker` shutting down is not unlikely. In such case the
`GossipEngine` should shut down as well.
With this patch a `<GossipEngine as Future>` finishes on
`Poll::Ready(None)` returned from `network_event_stream`.
* client/finality-grandpa/communication: Error on gossip engine finished
Have `<NetworkBridge as Future>::poll` return `Poll::Ready(Err)` instead
of `Poll::Ready(Ok)` to be consistent with the handling of the neighbor
packet worker stream and the gossip validator report stream. Both `Err`
as well as `Ok` shut down the `NetworkBridge` as well as the
`VoterWorker`.
* client/network-gossip/src/bridge: Add regression test
* client/network-gossip: Move substrate test client to dev dependencies
* client/network-gossip: Remove TODO
Addressed in a follow up pull request.
* client/network-gossip/bridge: Put match on newline after loop
* client/finality-grandpa/src/observer: Fix regression test
Make sure the event stream sender side is not dropped till the end.
* client/network-gossip: Merge GossipEngine and GossipEngineInner
Given that GossipEngine and GossipEngineInner are not shared between
threads anyone (public interface + background tasks), neither depends on
being Send or Sync. Thus one can merge the two as done in this patch.
One only needs to wrap an `Arc<Mutex<>>` around the whole structure when
the owner (e.g. finality-grandpa) needs to share the gossip engine
between threads.
* client/finality-grandpa: Wrap GossipEngine in Arc Mutex & lock it on use
GossipEngine in itself has no need to be Send and Sync, given that it
does not rely on separately spawned background tasks anymore. Given that
finality-grandpa shares the `NetworkBridge` potentially between threads
its components need to be clonable, thus this patch wraps `GossipEngine`
in an `Arc<Mutex<>>`.
* Open one substream for each notifications protocol
* Fix WASM build
* Apply suggestions from code review
Co-Authored-By: Toralf Wittner <tw@dtex.org>
* Address concerns
* Use unsigned-varint to read the varint
* Use unsigned-varint
* Forgot Cargo.lock
Co-authored-by: Toralf Wittner <tw@dtex.org>
`GossipEngine` spawns two tasks, one for a periodic tick, one to forward
messages from the network to subscribers. These tasks hold an `Arc` to a
`GossipEngineInner`.
To reduce the amount of shared ownership (locking) this patch integrates
the two tasks into a `Future` implementation on the `GossipEngine`
struct. This `Future` implementation can now be called from a single
owner, e.g. the `finality-grandpa` `NetworkBridge`.
As a side effect this removes the requirement on the `network-gossip`
crate to spawn tasks and thereby removes the requirement on the
`finality-grandpa` crate to spawn any tasks.
This is part of a greater effort to reduce the number of owners of
components within `finality-grandpa`, `network` and `network-gossip` as
well as to reduce the amount of unbounded channels. For details see
d9837d7dd, 5f80929dc and 597c0a6c4.
Aaro Altonen