Reward relayers for dispatching messages #385

svyatonik · 2020-09-29T10:05:54Z

related #316
follow up of #339 (II)

This PR introduces scheme which rewards relayers that have delivered messages to the target chain. There are no any transfers - reward logic is hidden behind MessageDeliveryAndDispatchPayment trait, so both "the winner takes all" and election scheme from #316 can be applied.

Main change is that inbound lane state now holds set of relayers which have delivered last messages. Since proof-of-delivery is that state, the set is included into the "delivery race" and is relayed back to the source chain.

The main drawback of this approach is that if there are too may relayed messages, this set may grow indefinitely. This is solved by introducing additional config param MaxUnconfirmedMessagesAtInboundLane. If relayers are not confirming delivery back to the source chain for too long, and there are MaxUnconfirmedMessagesAtInboundLane unconfirmed messages in the inbound lane, then it stops accepting new messages && waits until proof-of-outbound-lane-state will be included into the messages-proof. When updated outbound lane state is received, all rewarded relayers ids are removed from the relayers set. This would require minimal changes into "delivery race" logic of the relayer - I'll add those changes in separate PR.

HCastano

Looks fine, but I'd wait until Tomek takes a look before merging

HCastano · 2020-09-30T19:11:56Z

modules/message-lane/src/lib.rs

+				}
+
+				for message in lane_data.messages {
+					debug_assert_eq!(message.key.lane_id, lane_id);


Did you mean to keep this?

Can remove. But changing to something that'll work in release isn't desired, because this group_by is actually happens in the same pallet function.

primitives/message-lane/src/lib.rs

HCastano · 2020-09-30T19:27:45Z

primitives/message-lane/src/target_chain.rs

+use sp_std::{collections::btree_map::BTreeMap, fmt::Debug, prelude::*};
+
+/// Proved messages from the source chain.
+pub type ProvedMessages<Message> = BTreeMap<LaneId, ProvedLaneMessages<Message>>;


I'm unclear as to what a "proved" message is here. Is it one that's been proven to have been delivered, and thus one who's delivery can be acknowledged back on the source chain so we can reward the relayer?

When you relay messages to the target chain, they aren't simply encoded messages. Instead we relay storage-proof of these messages. ProvedMessages are the messages that are proved by that proof (of proved to be in that proof). If you have any other identifier idea - please suggest.

modules/message-lane/src/lib.rs

modules/message-lane/src/mock.rs

Co-authored-by: Hernando Castano <HCastano@users.noreply.github.com>

…y-bridges-common into reward-relayers

tomusdrw

Looks good, few concerns in comments.

modules/message-lane/src/inbound_lane.rs

primitives/message-lane/src/source_chain.rs

modules/message-lane/src/inbound_lane.rs

primitives/message-lane/src/target_chain.rs

modules/message-lane/src/lib.rs

Co-authored-by: Tomasz Drwięga <tomusdrw@users.noreply.github.com>

…y-bridges-common into reward-relayers

tomusdrw

lgtm!

modules/message-lane/src/lib.rs

* Add tests for checking messages above max limit Signed-off-by: MaciejBaj <macie.baj@gmail.com> * Extend the relayers entry of inbound lane by additional msg nonce Signed-off-by: MaciejBaj <macie.baj@gmail.com> * Support additional message nonce from inbound relayers Signed-off-by: MaciejBaj <macie.baj@gmail.com> * Code format Signed-off-by: MaciejBaj <macie.baj@gmail.com> * Merge messages range for highest relayers * Change unwrap() to ensure() while accessing relayers * Edit rustdocs for relayers deque at inbound lane data Co-authored-by: Hernando Castano <HCastano@users.noreply.github.com> * Declare additional relayers A & B and use across tests consistently Signed-off-by: MaciejBaj <macie.baj@gmail.com> * Remove duplicates and improve naming for inbound lane tests * Fix test checking max limit per inbound lane * Correct relayers rewards loop after a proof is received * Remove redundant check for messages ahead of received range * Correct grammar at inbound lane tests rustdocs Co-authored-by: Hernando Castano <HCastano@users.noreply.github.com> * Improve code quality of relayers updates 💅 Co-authored-by: Tomasz Drwięga <tomusdrw@users.noreply.github.com> * Test dispatches above max limit from same relayer Co-authored-by: Hernando Castano <HCastano@users.noreply.github.com> Co-authored-by: Tomasz Drwięga <tomusdrw@users.noreply.github.com>

* reward relayers for dispatching messages * clippy * Update modules/message-lane/src/lib.rs Co-authored-by: Hernando Castano <HCastano@users.noreply.github.com> * added comment * Update modules/message-lane/src/inbound_lane.rs Co-authored-by: Tomasz Drwięga <tomusdrw@users.noreply.github.com> * Update modules/message-lane/src/inbound_lane.rs Co-authored-by: Tomasz Drwięga <tomusdrw@users.noreply.github.com> * SubmitterId + RelayerId -> AccountId * add confirmation_relayer arg to pay_relayer_reward * cargo fmt --all * removed verify_and_decode_messages_proof from SourceHeaderChain * &mut self -> RefCell * Optimize max messages at inbound lane (paritytech#418) * Add tests for checking messages above max limit Signed-off-by: MaciejBaj <macie.baj@gmail.com> * Extend the relayers entry of inbound lane by additional msg nonce Signed-off-by: MaciejBaj <macie.baj@gmail.com> * Support additional message nonce from inbound relayers Signed-off-by: MaciejBaj <macie.baj@gmail.com> * Code format Signed-off-by: MaciejBaj <macie.baj@gmail.com> * Merge messages range for highest relayers * Change unwrap() to ensure() while accessing relayers * Edit rustdocs for relayers deque at inbound lane data Co-authored-by: Hernando Castano <HCastano@users.noreply.github.com> * Declare additional relayers A & B and use across tests consistently Signed-off-by: MaciejBaj <macie.baj@gmail.com> * Remove duplicates and improve naming for inbound lane tests * Fix test checking max limit per inbound lane * Correct relayers rewards loop after a proof is received * Remove redundant check for messages ahead of received range * Correct grammar at inbound lane tests rustdocs Co-authored-by: Hernando Castano <HCastano@users.noreply.github.com> * Improve code quality of relayers updates 💅 Co-authored-by: Tomasz Drwięga <tomusdrw@users.noreply.github.com> * Test dispatches above max limit from same relayer Co-authored-by: Hernando Castano <HCastano@users.noreply.github.com> Co-authored-by: Tomasz Drwięga <tomusdrw@users.noreply.github.com> * Fix typo. Co-authored-by: Hernando Castano <HCastano@users.noreply.github.com> Co-authored-by: Tomasz Drwięga <tomusdrw@users.noreply.github.com> Co-authored-by: Maciej Baj <macie.baj@gmail.com> Co-authored-by: Tomasz Drwięga <tomasz@parity.io>

reward relayers for dispatching messages

754233e

svyatonik added B-Substrate to Substrate P-Message Delivery labels Sep 29, 2020

clippy

0aa3735

svyatonik mentioned this pull request Sep 29, 2020

Integrate Rialto <-> Millau message lanes into Millau/Rialto runtimes #386

Merged

HCastano approved these changes Sep 30, 2020

View reviewed changes

HCastano requested a review from tomusdrw September 30, 2020 21:56

svyatonik and others added 3 commits October 2, 2020 09:32

Update modules/message-lane/src/lib.rs

a8b1e0d

Co-authored-by: Hernando Castano <HCastano@users.noreply.github.com>

added comment

fab0ffc

Merge branch 'reward-relayers' of https://github.com/paritytech/parit…

051f24d

…y-bridges-common into reward-relayers

tomusdrw reviewed Oct 6, 2020

View reviewed changes

tomusdrw mentioned this pull request Oct 6, 2020

Relayers Incentivisation #316

Closed

svyatonik and others added 6 commits October 6, 2020 17:54

Update modules/message-lane/src/inbound_lane.rs

0ea33b0

Co-authored-by: Tomasz Drwięga <tomusdrw@users.noreply.github.com>

Update modules/message-lane/src/inbound_lane.rs

d91b0e5

Co-authored-by: Tomasz Drwięga <tomusdrw@users.noreply.github.com>

SubmitterId + RelayerId -> AccountId

aa233de

Merge branch 'reward-relayers' of https://github.com/paritytech/parit…

f1e1990

…y-bridges-common into reward-relayers

add confirmation_relayer arg to pay_relayer_reward

0b62851

cargo fmt --all

3d28705

tomusdrw mentioned this pull request Oct 8, 2020

Optimize MaxUnconfirmedMessagesAtInboundLane #406

Closed

removed verify_and_decode_messages_proof from SourceHeaderChain

45310ad

MaciejBaj mentioned this pull request Oct 13, 2020

Optimize max messages at inbound lane #418

Merged

HCastano requested a review from tomusdrw October 14, 2020 18:33

Merge branch 'master' into reward-relayers

f9642cc

tomusdrw approved these changes Oct 19, 2020

View reviewed changes

modules/message-lane/src/lib.rs Outdated Show resolved Hide resolved

svyatonik and others added 3 commits October 19, 2020 15:44

&mut self -> RefCell

6222505

Fix typo.

efeae54

svyatonik merged commit 924d764 into master Oct 20, 2020

svyatonik deleted the reward-relayers branch October 20, 2020 10:12

svyatonik mentioned this pull request Oct 22, 2020

Optionally include outbound lane data into messages proof #428

Merged

svyatonik mentioned this pull request Oct 22, 2020

InboundLaneApi::latest_confirmed_nonce #429

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Reward relayers for dispatching messages #385

Reward relayers for dispatching messages #385

svyatonik commented Sep 29, 2020

HCastano left a comment

HCastano Sep 30, 2020

svyatonik Oct 2, 2020

HCastano Sep 30, 2020

svyatonik Oct 2, 2020

tomusdrw left a comment

tomusdrw left a comment

Reward relayers for dispatching messages #385

Reward relayers for dispatching messages #385

Conversation

svyatonik commented Sep 29, 2020

HCastano left a comment

Choose a reason for hiding this comment

HCastano Sep 30, 2020

Choose a reason for hiding this comment

svyatonik Oct 2, 2020

Choose a reason for hiding this comment

HCastano Sep 30, 2020

Choose a reason for hiding this comment

svyatonik Oct 2, 2020

Choose a reason for hiding this comment

tomusdrw left a comment

Choose a reason for hiding this comment

tomusdrw left a comment

Choose a reason for hiding this comment