Skip to main content

Pallet-dPoS for Parachain Staking

  • Proposer: Moonbeam Network
  • Payment Address: 0x9a66721302d9f30a9d11cf48a09c7ef1a842b5c8

Project Overview 📄

Moonbeam is an Ethereum-compatible parachain built with Substrate. Since January 2021, Moonbeam's runtime developer team has written and maintained a delegated Proof of Stake (dPoS) implementation designed specifically for parachains called parachain-staking. To make this code useful for other parachain teams, Moonbeam is applying for ecosystem funding to extend, benchmark, and document its dPoS implementation.

Background

In the Polkadot consensus model, parachains have different requirements and constraints than the relay chain. While frame/pallet-staking offers many features necessary for relay chain consensus and shared security, it is overkill for parachains, which operate with more limited execution resources.

Instead of running an on-chain election, parachain-staking implements direct delegation with a bounded number of nominations per AccountId (maximum of Config::MaxCollatorsPerNominator per account). In this paradigm, token holders (nominators) express exactly which collator candidates they would like to support and with what quantity of stake.

There is a new round every <Round<T>>::get().length = 600 blocks. At the start of each round, the top <TotalSelected<T>>::get() = 16 collator candidates (in terms of total backing stake) are chosen to be in the active collator set. At every block, a subset of this active collator set is chosen pseudorandomly by the author-filter pallet as valid block authors. Only valid block authors can produce blocks.

Block authors are logged via the set_author_inherent method in the author-inherent pallet. The logic in this runtime method reports each block author to parachain-staking, which distributes inflation rewards after a 2 round delay.

Team 👥

The PureStake team has extensive experience developing and delivering complex web2 software systems. In the last year, PureStake has built a number of web3 infrastructure projects including Algorand API Services (https://developer.purestake.io/) and the Goalseeker block explorer (https://goalseeker.purestake.io/), among others.

Derek co-founded Fuze in 2006 and as CTO was responsible for engineering, technical operations, product management, and design as the company grew over time into a 700 person cloud business. Alan was responsible for large parts of the Fuze cloud backend. Before joining Fuze, Alan was on the Google Streetview Team and also worked at NVidia developing drivers.

Antoine has been developing Web3 Dapps since 2016. He has participated in 3 projects inside ConsenSys (VariabL, Localties and Fathom) and helped HyperNetwork build their token system. Now, he's working for Moonbeam and has contributed to the polkadot-js suite (apps, common, api, etc).

Joshy and Amar worked at Parity on the Substrate Recipes.

Development Roadmap 🔩

  • Total Estimated Duration: 6 weeks
  • Full-time equivalent (FTE): 2 FTE
  • Total Costs: 30000 DAI

Both milestones will be worked on in parallel, but UI development will lag changes to the pallet implementation.

Milestone 1: Improve Parachain-Staking Pallet

  • Estimated Duration: 4 weeks
  • Costs: 20000 DAI

In addition to updating Rust crate-level docs and outdated user-level docs in step (5), we will also pursue the following changes to the current parachain-staking implementation:

  1. Benchmark to derive transaction weights. This work has been started in this PR.

  2. Security Research Labs reported two critical vulnerabilities in parachain-staking: (i) total locked balance was not updated in {collator, nominator}_bond_more leading to a potential underflow error (which could trigger excessive issuance) (ii) bounded number of nominations per collator allowed any account to fill the slots with the minimum nomination thereby preventing higher nominations.

  3. The inflation logic implemented in parachain-staking is minimal. Instead of integrating pallet-staking's reward curve, the current implementation calculates per-round inflation derived from an annual inflation rate. Although the inflation rate can be updated by governance (sudo as of now), it is constant. Some parachain teams (i.e. Kilt) have requested configurable inflation that uses pallet-staking's reward curve instead because it has been audited and reviewed more closely.

  4. Moonbeam reserves 30% of inflation for future parachain bond(s). To support this functionality, parachain-staking added the storage item ParachainBondConfig. This storage item is updatable by the root origin; it configures the percent (30%) of inflation reserved as well as the AccountId which receives the reserved funds. This feature is convenient for parachains in the Polkadot ecosystem, all of which must pay rent to the network by locking funds in the parachain bond.

NumberDeliverableSpecification
0a.LicenseGPL3
0d.DocumentationUpdate Rust crate-level docs and user-level docs with implementation details.
1.BenchmarkBenchmark the existing implementation to derive transaction weights and discern optimization opportunities.
2.Patch SR-Labs Critical VulnerabilitiesPatch both critical vulnerabilities reported by SR labs.
3.Configurable InflationReplace sudo with governance origin for setting inflation rate. Provide instructions for replacing constant inflation with pallet-staking's reward curve logic.
4.Configurable Parachain Bond ReservationAdd optional parachain bond configuration that enables reserving a portion of inflation for future parachain bonds.

Milestone 2: Parachain-Staking Polkadot-JS UI

  • Estimated Duration: 4 weeks
  • Costs: 10000 DAI
NumberDeliverableSpecification
0a.LicenseApache 2.0
1.Custom Polkadot-JS UIAn overlay UI using polkadot-js similar to the pallet-staking UI
2.PR polkadot-js appsMake a pull request to polkadot-js apps with output

Additional Information ➕

The current implementation has significant unit test coverage and this will be maintained throughout the changes listed in Milestone 1. There are also runtime integration tests written in Rust as well as TS integration tests. Maintenance is not explicitly included in the milestones because test coverage is already relatively comprehensive and we have an incentive to maintain it throughout the proposed changes.