Indexing Rewards and The Merge

Governance & GIPs Graph Improvement Proposals (GIP)
1

I want to bring attention about something. Ethereum is merging to the PoS chain during September. At the application level it doesn’t involve any change to the contracts but the blockTime will change from average of 13.4s to 12s fixed slots.

The rewards mechanism is currently using an issuance per block value, that means that now the clock will run faster and issuance speed will increase on the PoS chain.

Current value:
issuanceRate: "1000000012184945188" # per block increase of total supply, blocks in a year = 365*60*60*24/13

That gets us to an effective issuance rate of 3% a year when compounding the rate to the numebr of blocks.

Current value: (after-merge)

After The Merge we will have 2628000 blocks a year instead of 2425846, which at current per block issuance rate it will lead to yearly 3.2% rate.

The effective given indexing rewards during the first year was close to 2.7% due to the effect of opted-out rewards in allocations with POI=0.

It is up to the Council to decide if to adjust the issuance rate to make it more precise to the Ethereum PoS chain.

8 Likes
2

Would like to have community input on this too, but putting this on the agenda for the Council bi-weekly meeting on Thursday.

3 Likes
3

Fully support this. Changes to the inflation rate should be an explicit parameter change, rather than something we inherit from a change in Ethereum.

3 Likes
4

I tested the issuance rate parameter change using a Mainnet fork, simulated the change and ran a monitor script to check that pending rewards looks good for every allocation after the change.

I’m going to initiate this change in the Goerli testnet and after a day running, it will be ready for the Council to consider initiating on Mainnet.

6 Likes
5

I’m strongly in favor of submitting a protocol parameter change to keep issuance at 3%.

I don’t believe The Graph Council should permit any changes to the 3% issuance rate, even through inaction, until rigorous research has been done on what a better long-term issuance rate might be.

Allowing the issuance rate to increase past 3% would set a poor precedent.

2 Likes
6

On a somewhat related note, we might also want to adjust the epoch length from the current 6,646 blocks to 7,200 (12s blocks). Chatted with @ari and the implications of that are much less meaningful (undelegation/unstake is a bit faster), but we should consider that change too

3 Likes
7

Hey all, posting it here for reference.

The transaction to update the indexing rewards issuance rate to follow the ETH post-merge block speed was executed by the Council in transaction Ethereum Transaction Hash (Txhash) Details | Etherscan and is up since block 15841198.

This was part of GGP-0013 voted in Snapshot by The Graph Council.

3 Likes