Integrate litep2p into Polkadot SDK (#2944)

[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>

polkadot/node/network/protocol/src/authority_discovery.rs

@@ -23,7 +23,8 @@ use async_trait::async_trait;
 use sc_authority_discovery::Service as AuthorityDiscoveryService;
 
 use polkadot_primitives::AuthorityDiscoveryId;
-use sc_network::{Multiaddr, PeerId};
+use sc_network::Multiaddr;
+use sc_network_types::PeerId;
 
 /// An abstraction over the authority discovery service.
 ///

polkadot/node/network/protocol/src/lib.rs

@@ -25,7 +25,8 @@ use std::{collections::HashMap, fmt};
 
 #[doc(hidden)]
 pub use polkadot_node_jaeger as jaeger;
-pub use sc_network::{IfDisconnected, PeerId};
+pub use sc_network::IfDisconnected;
+pub use sc_network_types::PeerId;
 #[doc(hidden)]
 pub use std::sync::Arc;
 
@@ -610,7 +611,7 @@ pub mod v1 {
 	///
 	/// The payload is the local peer id of the node, which serves to prove that it
 	/// controls the collator key it is declaring an intention to collate under.
-	pub fn declare_signature_payload(peer_id: &sc_network::PeerId) -> Vec<u8> {
+	pub fn declare_signature_payload(peer_id: &sc_network_types::PeerId) -> Vec<u8> {
 		let mut payload = peer_id.to_bytes();
 		payload.extend_from_slice(b"COLL");
 		payload
@@ -863,7 +864,7 @@ pub mod v2 {
 	///
 	/// The payload is the local peer id of the node, which serves to prove that it
 	/// controls the collator key it is declaring an intention to collate under.
-	pub fn declare_signature_payload(peer_id: &sc_network::PeerId) -> Vec<u8> {
+	pub fn declare_signature_payload(peer_id: &sc_network_types::PeerId) -> Vec<u8> {
 		let mut payload = peer_id.to_bytes();
 		payload.extend_from_slice(b"COLL");
 		payload

polkadot/node/network/protocol/src/peer_set.rs

@@ -19,13 +19,14 @@
 use derive_more::Display;
 use polkadot_primitives::Hash;
 use sc_network::{
-	config::{NonDefaultSetConfig, SetConfig},
-	types::ProtocolName,
-	NotificationService,
+	config::SetConfig, peer_store::PeerStoreProvider, service::NotificationMetrics,
+	types::ProtocolName, NetworkBackend, NotificationService,
 };
+use sp_runtime::traits::Block;
 use std::{
 	collections::{hash_map::Entry, HashMap},
 	ops::{Index, IndexMut},
+	sync::Arc,
 };
 use strum::{EnumIter, IntoEnumIterator};
 
@@ -65,11 +66,13 @@ impl PeerSet {
 	///
 	/// Those should be used in the network configuration to register the protocols with the
 	/// network service.
-	pub fn get_info(
+	pub fn get_info<B: Block, N: NetworkBackend<B, <B as Block>::Hash>>(
 		self,
 		is_authority: IsAuthority,
 		peerset_protocol_names: &PeerSetProtocolNames,
-	) -> (NonDefaultSetConfig, (PeerSet, Box<dyn NotificationService>)) {
+		metrics: NotificationMetrics,
+		peer_store_handle: Arc<dyn PeerStoreProvider>,
+	) -> (N::NotificationProtocolConfig, (PeerSet, Box<dyn NotificationService>)) {
 		// Networking layer relies on `get_main_name()` being the main name of the protocol
 		// for peersets and connection management.
 		let protocol = peerset_protocol_names.get_main_name(self);
@@ -82,7 +85,7 @@ impl PeerSet {
 
 		match self {
 			PeerSet::Validation => {
-				let (config, notification_service) = NonDefaultSetConfig::new(
+				let (config, notification_service) = N::notification_config(
 					protocol,
 					fallback_names,
 					max_notification_size,
@@ -97,12 +100,14 @@ impl PeerSet {
 						reserved_nodes: Vec::new(),
 						non_reserved_mode: sc_network::config::NonReservedPeerMode::Accept,
 					},
+					metrics,
+					peer_store_handle,
 				);
 
 				(config, (PeerSet::Validation, notification_service))
 			},
 			PeerSet::Collation => {
-				let (config, notification_service) = NonDefaultSetConfig::new(
+				let (config, notification_service) = N::notification_config(
 					protocol,
 					fallback_names,
 					max_notification_size,
@@ -119,6 +124,8 @@ impl PeerSet {
 							sc_network::config::NonReservedPeerMode::Deny
 						},
 					},
+					metrics,
+					peer_store_handle,
 				);
 
 				(config, (PeerSet::Collation, notification_service))
@@ -207,12 +214,21 @@ impl<T> IndexMut<PeerSet> for PerPeerSet<T> {
 ///
 /// Should be used during network configuration (added to `NetworkConfiguration::extra_sets`)
 /// or shortly after startup to register the protocols with the network service.
-pub fn peer_sets_info(
+pub fn peer_sets_info<B: Block, N: NetworkBackend<B, <B as Block>::Hash>>(
 	is_authority: IsAuthority,
 	peerset_protocol_names: &PeerSetProtocolNames,
-) -> Vec<(NonDefaultSetConfig, (PeerSet, Box<dyn NotificationService>))> {
+	metrics: NotificationMetrics,
+	peer_store_handle: Arc<dyn PeerStoreProvider>,
+) -> Vec<(N::NotificationProtocolConfig, (PeerSet, Box<dyn NotificationService>))> {
 	PeerSet::iter()
-		.map(|s| s.get_info(is_authority, &peerset_protocol_names))
+		.map(|s| {
+			s.get_info::<B, N>(
+				is_authority,
+				&peerset_protocol_names,
+				metrics.clone(),
+				Arc::clone(&peer_store_handle),
+			)
+		})
 		.collect()
 }
 

polkadot/node/network/protocol/src/request_response/incoming/error.rs

@@ -16,7 +16,7 @@
 
 //! Error handling related code and Error/Result definitions.
 
-use sc_network::PeerId;
+use sc_network_types::PeerId;
 
 use parity_scale_codec::Error as DecodingError;
 

polkadot/node/network/protocol/src/request_response/incoming/mod.rs

@@ -20,7 +20,9 @@ use futures::{channel::oneshot, StreamExt};
 
 use parity_scale_codec::{Decode, Encode};
 
-use sc_network::{config as netconfig, config::RequestResponseConfig, PeerId};
+use sc_network::{config as netconfig, NetworkBackend};
+use sc_network_types::PeerId;
+use sp_runtime::traits::Block;
 
 use super::{IsRequest, ReqProtocolNames};
 use crate::UnifiedReputationChange;
@@ -52,10 +54,10 @@ where
 	///
 	/// This Register that config with substrate networking and receive incoming requests via the
 	/// returned `IncomingRequestReceiver`.
-	pub fn get_config_receiver(
+	pub fn get_config_receiver<B: Block, N: NetworkBackend<B, <B as Block>::Hash>>(
 		req_protocol_names: &ReqProtocolNames,
-	) -> (IncomingRequestReceiver<Req>, RequestResponseConfig) {
-		let (raw, cfg) = Req::PROTOCOL.get_config(req_protocol_names);
+	) -> (IncomingRequestReceiver<Req>, N::RequestResponseProtocolConfig) {
+		let (raw, cfg) = Req::PROTOCOL.get_config::<B, N>(req_protocol_names);
 		(IncomingRequestReceiver { raw, phantom: PhantomData {} }, cfg)
 	}
 

polkadot/node/network/protocol/src/request_response/mod.rs

@@ -52,6 +52,8 @@
 use std::{collections::HashMap, time::Duration, u64};
 
 use polkadot_primitives::{MAX_CODE_SIZE, MAX_POV_SIZE};
+use sc_network::NetworkBackend;
+use sp_runtime::traits::Block;
 use strum::{EnumIter, IntoEnumIterator};
 
 pub use sc_network::{config as network, config::RequestResponseConfig, ProtocolName};
@@ -179,76 +181,76 @@ impl Protocol {
 	///
 	/// Returns a `ProtocolConfig` for this protocol.
 	/// Use this if you plan only to send requests for this protocol.
-	pub fn get_outbound_only_config(
+	pub fn get_outbound_only_config<B: Block, N: NetworkBackend<B, <B as Block>::Hash>>(
 		self,
 		req_protocol_names: &ReqProtocolNames,
-	) -> RequestResponseConfig {
-		self.create_config(req_protocol_names, None)
+	) -> N::RequestResponseProtocolConfig {
+		self.create_config::<B, N>(req_protocol_names, None)
 	}
 
 	/// Get a configuration for a given Request response protocol.
 	///
 	/// Returns a receiver for messages received on this protocol and the requested
 	/// `ProtocolConfig`.
-	pub fn get_config(
+	pub fn get_config<B: Block, N: NetworkBackend<B, <B as Block>::Hash>>(
 		self,
 		req_protocol_names: &ReqProtocolNames,
-	) -> (async_channel::Receiver<network::IncomingRequest>, RequestResponseConfig) {
+	) -> (async_channel::Receiver<network::IncomingRequest>, N::RequestResponseProtocolConfig) {
 		let (tx, rx) = async_channel::bounded(self.get_channel_size());
-		let cfg = self.create_config(req_protocol_names, Some(tx));
+		let cfg = self.create_config::<B, N>(req_protocol_names, Some(tx));
 		(rx, cfg)
 	}
 
-	fn create_config(
+	fn create_config<B: Block, N: NetworkBackend<B, <B as Block>::Hash>>(
 		self,
 		req_protocol_names: &ReqProtocolNames,
 		tx: Option<async_channel::Sender<network::IncomingRequest>>,
-	) -> RequestResponseConfig {
+	) -> N::RequestResponseProtocolConfig {
 		let name = req_protocol_names.get_name(self);
 		let legacy_names = self.get_legacy_name().into_iter().map(Into::into).collect();
 		match self {
-			Protocol::ChunkFetchingV1 => RequestResponseConfig {
+			Protocol::ChunkFetchingV1 => N::request_response_config(
 				name,
-				fallback_names: legacy_names,
-				max_request_size: 1_000,
-				max_response_size: POV_RESPONSE_SIZE as u64 * 3,
+				legacy_names,
+				1_000,
+				POV_RESPONSE_SIZE as u64 * 3,
 				// We are connected to all validators:
-				request_timeout: CHUNK_REQUEST_TIMEOUT,
-				inbound_queue: tx,
-			},
+				CHUNK_REQUEST_TIMEOUT,
+				tx,
+			),
 			Protocol::CollationFetchingV1 | Protocol::CollationFetchingV2 =>
-				RequestResponseConfig {
+				N::request_response_config(
 					name,
-					fallback_names: legacy_names,
-					max_request_size: 1_000,
-					max_response_size: POV_RESPONSE_SIZE,
+					legacy_names,
+					1_000,
+					POV_RESPONSE_SIZE,
 					// Taken from initial implementation in collator protocol:
-					request_timeout: POV_REQUEST_TIMEOUT_CONNECTED,
-					inbound_queue: tx,
-				},
-			Protocol::PoVFetchingV1 => RequestResponseConfig {
+					POV_REQUEST_TIMEOUT_CONNECTED,
+					tx,
+				),
+			Protocol::PoVFetchingV1 => N::request_response_config(
 				name,
-				fallback_names: legacy_names,
-				max_request_size: 1_000,
-				max_response_size: POV_RESPONSE_SIZE,
-				request_timeout: POV_REQUEST_TIMEOUT_CONNECTED,
-				inbound_queue: tx,
-			},
-			Protocol::AvailableDataFetchingV1 => RequestResponseConfig {
+				legacy_names,
+				1_000,
+				POV_RESPONSE_SIZE,
+				POV_REQUEST_TIMEOUT_CONNECTED,
+				tx,
+			),
+			Protocol::AvailableDataFetchingV1 => N::request_response_config(
 				name,
-				fallback_names: legacy_names,
-				max_request_size: 1_000,
+				legacy_names,
+				1_000,
 				// Available data size is dominated by the PoV size.
-				max_response_size: POV_RESPONSE_SIZE,
-				request_timeout: POV_REQUEST_TIMEOUT_CONNECTED,
-				inbound_queue: tx,
-			},
-			Protocol::StatementFetchingV1 => RequestResponseConfig {
+				POV_RESPONSE_SIZE,
+				POV_REQUEST_TIMEOUT_CONNECTED,
+				tx,
+			),
+			Protocol::StatementFetchingV1 => N::request_response_config(
 				name,
-				fallback_names: legacy_names,
-				max_request_size: 1_000,
+				legacy_names,
+				1_000,
 				// Available data size is dominated code size.
-				max_response_size: STATEMENT_RESPONSE_SIZE,
+				STATEMENT_RESPONSE_SIZE,
 				// We need statement fetching to be fast and will try our best at the responding
 				// side to answer requests within that timeout, assuming a bandwidth of 500Mbit/s
 				// - which is the recommended minimum bandwidth for nodes on Kusama as of April
@@ -258,27 +260,27 @@ impl Protocol {
 				// waiting for timeout on an overloaded node.  Fetches from slow nodes will likely
 				// fail, but this is desired, so we can quickly move on to a faster one - we should
 				// also decrease its reputation.
-				request_timeout: Duration::from_secs(1),
-				inbound_queue: tx,
-			},
-			Protocol::DisputeSendingV1 => RequestResponseConfig {
+				Duration::from_secs(1),
+				tx,
+			),
+			Protocol::DisputeSendingV1 => N::request_response_config(
 				name,
-				fallback_names: legacy_names,
-				max_request_size: 1_000,
+				legacy_names,
+				1_000,
 				// Responses are just confirmation, in essence not even a bit. So 100 seems
 				// plenty.
-				max_response_size: 100,
-				request_timeout: DISPUTE_REQUEST_TIMEOUT,
-				inbound_queue: tx,
-			},
-			Protocol::AttestedCandidateV2 => RequestResponseConfig {
+				100,
+				DISPUTE_REQUEST_TIMEOUT,
+				tx,
+			),
+			Protocol::AttestedCandidateV2 => N::request_response_config(
 				name,
-				fallback_names: legacy_names,
-				max_request_size: 1_000,
-				max_response_size: ATTESTED_CANDIDATE_RESPONSE_SIZE,
-				request_timeout: ATTESTED_CANDIDATE_TIMEOUT,
-				inbound_queue: tx,
-			},
+				legacy_names,
+				1_000,
+				ATTESTED_CANDIDATE_RESPONSE_SIZE,
+				ATTESTED_CANDIDATE_TIMEOUT,
+				tx,
+			),
 		}
 	}
 

polkadot/node/network/protocol/src/request_response/outgoing.rs

@@ -20,7 +20,7 @@ use network::ProtocolName;
 use parity_scale_codec::{Decode, Encode, Error as DecodingError};
 
 use sc_network as network;
-use sc_network::PeerId;
+use sc_network_types::PeerId;
 
 use polkadot_primitives::AuthorityDiscoveryId;