Allow for a dynamic number of nominators #10340

kianenigma · 2021-11-22T18:27:31Z

This PR adds two flexible pieces of logic to pallet-staking:

The number of votes that each nominator can cast can be configurable now. A simple implementation would use a constant value for this. A fancier implementation would make this a function of the amount of stake.
The amount of voters that is taken for the election snapshot is now dynamic over two axis. A SnapshotBound is defined as a cap on both the count and byte-size of the snapshot. The snapshot creation will, if bounded, stop whenever either of the bounds are met.

Breaking Changes

MAX_NOMINATIONS is removed from the API of pallet-staking and is replaced by NominationQuota. To replicate the same behaviour, use pallet_staking::FixedNomintionQuota<OLD_VALUE>.

const OLD_MAX_NOMINATIONS: u32 = 16;
impl pallet_staking::Config {
  ... 
  type NominationQuota = pallet_staking::FixedNomintionQuota<OLD_MAX_NOMINATIONS >;
}

VoterSnapshotPerBlock is removed. Two replacements are introduced, VoterSnapshotBounds and TargetSnapshotBounds. To replicate exactly the old behaviour, VoterSnapshotBounds should be SnapshotBounds::new_count(OLD_VALUE) and TargetSnapshotBounds should be SnapshotBounds::new_unbounded(). Note that this is only safe when the number of validators in staking is well bounded (MaxValidatorsCount is set to some small value). As seen in this PR (node/runtime/src/lib.rs), it is recommended to replace this with a small sane bound.

…e-max-nomination-dynamic

kianenigma · 2021-11-24T09:36:52Z

For the sake of prosperity, I hereby confirm that commit 2f629a9 which is not properly signed was made by me.

kianenigma · 2021-11-29T12:40:35Z

frame/staking/src/lib.rs

+///  Essentially, draws a line between `(MinBalance, MinQuota)` and `(MaxBalance, MaxQuota)`, Also,
+/// values less than `MinBalance` have `MinQuota` and values higher than `MaxBalance` have
+/// `MaxQuota`.
+pub struct LinearNominationQuota<MinBalance, MaxBalance, const MAX_QUOTA: u32, const MIN_QUOTA: u32>(


I won't test this further yet because I am not sure if we are going to use it.

kianenigma · 2021-11-29T13:14:46Z

frame/staking/src/pallet/impls.rs

+/// Whilst doing this, [`size`] will track the entire size of the `Vec<Voter>`, except for the
+/// length prefix of the outer `Vec`. To get the final size at any point, use
+/// [`final_byte_size_of`].
+pub(crate) struct StaticSizeTracker<AccountId> {


I would make this fully generic and reusable, but it is not really feasible given the nested vec.

jonasW3F · 2021-11-30T18:26:05Z

Based on the discussion in the element group, I'd like to summarize my points on this. First, I think reducing the number of nominations per nominator is generally a reasonable approach if we can trade those for increasing the number of total nominators in the short run. But, I think it is crucial to not reduce those too far, because for the following reasons:

The most important reason is that we want our nominators to participate in decentralizing the network, which means that they should seek out the best validators to fit their preferences and include them in their favorite set. Also, we would like them to include inactive validators because of the mentioned reason. Reducing the nominations too much would jeopardize this. Especially including inactive validators becomes increasingly costly which favors large operators and reduces healthy turnover in the active set.
Nominators could frequently miss out on rewards if the number of nominations becomes too low, which would cause frustration.
I think users of the network generally enjoy the freedom of choice that NPoS gives them and it is an important feature that distinguishes us from other networks that restrict the choices of their stakers.
Once we obtain better scaling solutions for staking, sacrificing on the first three mentioned arguments would hurt the network in the long run.

Having said that, I think there is still room to reduce the nominations (currently 16 in Polkadot) to free space for more nominators. As mentioned in the group, I'd propose a rule of setting the minimum number of nominations to 6-8 and giving another nomination per 50k DOT in the stash. That would provide sufficient freedom to smaller stakers as well as for larger token holders.

This strategy of 6 + 1 every 50k (and cap at 32) would give, based on some recent era data, a maximum of 111801 nominations. If we assume that nominators use their free slots similarly as currently (64%), that would bring down the number of nominations to 72k. If we compare that to the current number of 184954 nominations, that's a good gain.

kianenigma · 2022-01-03T16:59:23Z

I'm not quite sure about the prospect of this, so marking it is on ice for now until I know more, or I will temporarily close it.

kianenigma · 2022-03-07T09:57:57Z

frame/election-provider-support/src/lib.rs

@@ -439,3 +443,173 @@ impl<
 		sp_npos_elections::phragmms(winners, targets, voters, Balancing::get())
 	}
 }
+
+/// The limits imposed on a snapshot, either voters or targets.
+#[derive(Clone, Copy, RuntimeDebug, scale_info::TypeInfo, Encode, Decode)]


This is a good option to be backported back to master in either case.

stale · 2022-05-06T16:12:06Z

Hey, is anyone still working on this? Due to the inactivity this issue has been automatically marked as stale. It will be closed if no further activity occurs. Thank you for your contributions.

kianenigma · 2022-05-06T17:56:17Z

YES

kianenigma · 2022-05-19T09:37:02Z

too stale, giving up for now

kianenigma · 2022-07-06T11:46:32Z

This is back from the grave as per a discussion with @xlc, which led us to re-discover this as a useful feature. I think it is a great feature to add, and we can later let the governance decide what the range of nomination should be. We can and should deploy this initially in a backwards compatible way, i.e. with FixedNomintionQuota.

gpestana · 2022-12-12T21:50:15Z

I think the logic and design of the PR look good, but the git history has diverged quite significantly from the current master. My plans to revive this PR are to cut a new branch from master and:

cherry-pick and resolve conflicts of (all in this PR)
- 12872c79ed2d26b460952da6035fcb3d002609ec
- f9a27f1b7809434d6bf653aea1021a4a76ee26aa
- 9844424cd49eb84d95d878070dffa7e0edbbb8e4
- 7b4e7a8f3a3f34a302785f7dcf0085d69b3b1488
redesign the exhausts_size_count_non_zero api to make sure that the caller passes given_size and given_count in the correct order
update benchmarks
revisit tests
open companion PRs

cc @kianenigma

Polkadot-Forum · 2023-01-30T21:23:00Z

This pull request has been mentioned on Polkadot Forum. There might be relevant details there:

https://forum.polkadot.network/t/the-future-of-polkadot-staking/1848/6

Polkadot-Forum · 2023-01-30T22:18:32Z

This pull request has been mentioned on Polkadot Forum. There might be relevant details there:

https://forum.polkadot.network/t/the-future-of-polkadot-staking/1848/7

Polkadot-Forum · 2023-02-03T16:58:42Z

This pull request has been mentioned on Polkadot Forum. There might be relevant details there:

https://forum.polkadot.network/t/the-future-of-polkadot-staking/1848/8

kianenigma added 5 commits November 12, 2021 15:07

first draft

12872c7

new version with dual measurement

f9a27f1

fix a bunch of tests as well

9844424

Merge branch 'master' of github.com:paritytech/substrate into kiz-mor…

151060c

…e-max-nomination-dynamic

it all works now

7b4e7a8

kianenigma added the A3-in_progress Pull request is in progress. No review needed at this stage. label Nov 22, 2021

kianenigma and others added 3 commits November 23, 2021 10:22

Merge branch 'master' of github.com:paritytech/substrate into kiz-mor…

d8297e1

…e-max-nomination-dynamic

A test commit from the new macbook

2f629a9

new test commit to enable signing

a60ca9f

kianenigma commented Nov 29, 2021

View reviewed changes

Master.into()

b7d5c2a

kianenigma marked this pull request as ready for review November 30, 2021 07:52

kianenigma requested a review from andresilva as a code owner November 30, 2021 07:52

github-actions bot added A0-please_review Pull request needs code review. and removed A3-in_progress Pull request is in progress. No review needed at this stage. labels Nov 30, 2021

kianenigma added B7-runtimenoteworthy C1-low PR touches the given topic and has a low impact on builders. D9-needsaudit 👮 PR contains changes to fund-managing logic that should be properly reviewed and externally audited labels Nov 30, 2021

remove election playground stuff'

95bbaf2

Master.into()

6c9c0fd

kianenigma added A1-onice and removed A0-please_review Pull request needs code review. D9-needsaudit 👮 PR contains changes to fund-managing logic that should be properly reviewed and externally audited labels Jan 3, 2022

kianenigma mentioned this pull request Mar 7, 2022

lower kusama staking limits paritytech/polkadot#5000

Merged

kianenigma commented Mar 7, 2022

View reviewed changes

stale bot added the A5-stale Pull request did not receive any updates in a long time. No review needed at this stage. Close it. label Apr 6, 2022

paritytech deleted a comment from stale bot Apr 6, 2022

stale bot removed the A5-stale Pull request did not receive any updates in a long time. No review needed at this stage. Close it. label Apr 6, 2022

stale bot added the A5-stale Pull request did not receive any updates in a long time. No review needed at this stage. Close it. label May 6, 2022

stale bot removed the A5-stale Pull request did not receive any updates in a long time. No review needed at this stage. Close it. label May 6, 2022

kianenigma closed this May 19, 2022

kianenigma reopened this Jul 6, 2022

kianenigma closed this Jul 13, 2022

kianenigma reopened this Dec 12, 2022

kianenigma closed this Dec 16, 2022

This was referenced Dec 19, 2022

Allow for a dynamic number of nominators #12968

Closed

[NPoS] Implements dynamic number of nominators #12970

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Allow for a dynamic number of nominators #10340

Allow for a dynamic number of nominators #10340

kianenigma commented Nov 22, 2021 •

edited

Loading

kianenigma commented Nov 24, 2021

kianenigma Nov 29, 2021

kianenigma Nov 29, 2021

jonasW3F commented Nov 30, 2021

kianenigma commented Jan 3, 2022

kianenigma Mar 7, 2022

stale bot commented May 6, 2022

kianenigma commented May 6, 2022

kianenigma commented May 19, 2022

kianenigma commented Jul 6, 2022 •

edited

Loading

gpestana commented Dec 12, 2022 •

edited

Loading

Polkadot-Forum commented Jan 30, 2023

Polkadot-Forum commented Jan 30, 2023

Polkadot-Forum commented Feb 3, 2023

Allow for a dynamic number of nominators #10340

Allow for a dynamic number of nominators #10340

Conversation

kianenigma commented Nov 22, 2021 • edited Loading

Breaking Changes

kianenigma commented Nov 24, 2021

kianenigma Nov 29, 2021

Choose a reason for hiding this comment

kianenigma Nov 29, 2021

Choose a reason for hiding this comment

jonasW3F commented Nov 30, 2021

kianenigma commented Jan 3, 2022

kianenigma Mar 7, 2022

Choose a reason for hiding this comment

stale bot commented May 6, 2022

kianenigma commented May 6, 2022

kianenigma commented May 19, 2022

kianenigma commented Jul 6, 2022 • edited Loading

gpestana commented Dec 12, 2022 • edited Loading

Polkadot-Forum commented Jan 30, 2023

Polkadot-Forum commented Jan 30, 2023

Polkadot-Forum commented Feb 3, 2023

kianenigma commented Nov 22, 2021 •

edited

Loading

kianenigma commented Jul 6, 2022 •

edited

Loading

gpestana commented Dec 12, 2022 •

edited

Loading