In early June, Mike Garland from node operator Alchemy gave a 1.5 hour presentation on Ethereum 2.0 staking, hosted by Jehan Chu of Kenetic Capital. I thought it was interesting so I took some notes. Errors and omissions are my own.
Eth 2.0 rollout
- Eth 2.0 Rollout will be in 3 phases:
- Phase 0 – Beacon Chain – July 2020
- Validators and basic Proof of Stake functionality
- No smart contracts, no accounts
- Staking rewards start when the threshold number of validators are reached
- Phase 1 – Sharding – 2021
- Addition of 64 “Shard Chains” for scalability
- Phase 2 – Accounts and contracts – 2022
- Remainder of the “normal” Ethereum functionality – smart contracts, etc.
- Phase 0 – Beacon Chain – July 2020
Promises of Eth 2.0
- Promises of Eth 2.0
- Scalability (send more transactions through the network without paying tons of gas or waiting a long time)
- Sharding is the main mechanism.
- Eth 1.0 – 14tx/s and getting slower and more expensive
- Eth 2.0 – up to 7,500tx/s and scales with the number of shards
- (Visa processes 1,700tx/s)
- Reminder: Phase 0 of Eth 2.0 will have 0 tx/s. No transactions. Initial beacon chain is for staking only. No transfers, state changes, or smart contracts. Eth 2.0 only becomes “useable as expected” in Phase 2.
- Accessibility (make sure the features are available to all users, not just miners or those with lots of compute power)
- Security (greater security of the chain, related to staking decentralisation)
- Scalability (send more transactions through the network without paying tons of gas or waiting a long time)
Proof of Stake vs Proof of Work
- Proof of Work mining is more energy hungry than gold digging: 14 MJ/$1 of BTC vs 5MJ/$1 of gold!
- Proof of Stake is more energy efficient (no mining), in theory more decentralised (no specialised hardware required), more secure (more diverse set of miners so 51% attacks are more unlikely) than Proof of Work.
The role of validators
- Validators
- Vote on the validity of the next block (called “attesting”)
- Can propose new blocks to be voted on
- Can earn rewards for successful participation in the network
- Can be punished in two ways:
- Slashing: Large fines, and kicked out, for appearing to attempt to cheat (whether accidentally or with intent to cheat)
- eg voting for multiple blocks at the same height
- Penalties: Small fines for downtime / unresponsiveness
- Slashing: Large fines, and kicked out, for appearing to attempt to cheat (whether accidentally or with intent to cheat)
Validator lifecycle
- Validator lifecycle: Initial deposit → Pending → Active → Exiting or Slashed → Exited.
- Spend most of your time in “Active” state: Proposing on blocks, voting on blocks etc
- Initial Deposit
- Deposit of ETH is made to a specific contract on Eth 1.0 Network.
- Validator remains in “depositing” state for approx 7.5 hours to avoid any potential of block reorgs (due to probabilistic nature of Eth 1.0 PoW).
- Pending State
- Deposit is recognised by Eth 2.0 chain
- Validator is officially recognised by Eth 2.0
- If the deposit is at least 32 ETH, validator joins a queue / waiting list. Mininum wait estimated at 25 mins if the queue is empty, up to days/weeks if the queue is full. This helps slow entry/exit of validators.
- Eth 2.0 wants to have a stable validator set, so only small number of validators can start/stop at one time.
- At the front of the queue, the validator can become active.
- Active State
- Validators spend most of their time here.
- As an active validator, you must attest to blocks, else there are (small) penalties.
- Attest at least once every 6 minutes (once every epic)
- Occasionally can propose blocks if you want.
- Validator will remain active unless:
- Funds drop below 16 ETH
- Validator asks the network to stop (voluntary exit)
- Validator is caught cheating (specific types of cheating)
- Exiting State
- Validator joins a queue to exit
- Exiting validators must continue to behave as active until exited, and are still subject to penalties while in the exit queue.
- Validators cannot stop attesting immediately, without penalty
- Reduces “hit and run” types of fraud
- Slashed State
- Occurs if a validator is caught cheating
- Validator is immediately fined 1 ETH
- Labelled as a cheater, forced to exit and marked for additional penalties
- Slashing can occur even when exiting, or in retrospect when exited
- Exited State
- End of lifecycle, no more need to attest or propose
- 1 day delay before funds can be collected
- If validator was slashed, Fund delay is increased to 36 days, and after 18 days, validator will be penalised an additional amount based on how many other validators were slashed.
- Exited for Slashing
- If 1/3 of the validator set are slashed at the same time you lose 100% of your stake. If it’s just you, it’s like 4% or something then straight line up to 100% at 33% of validator set.
- Note: this is counterintuitive so will say again. The more validators that are implicated in a cheating event, the more each validator is punished. The opposite of “safety in numbers”.
- If 1/3 of the validator set are slashed at the same time you lose 100% of your stake. If it’s just you, it’s like 4% or something then straight line up to 100% at 33% of validator set.
- Slashed funds redistributed to other “healthy” validators (at a guess based on other staking mechanisms)
- Penalties – differ under circumstance. Eg if multiple validators go offline at once, they’re punished more than if only one goes offline. Attempt to prevent centralisation of technologies, centralisation of clouds etc. You are incentivised to run on a different cloud provider vs others.
Minimum requirements to run a validator
- One validator one vote (not one coin one vote, which is what many other PoS protocols do)
- 32 ETH is the minimum stake to be considered a valid validator.
- Your income from validating is based on the number of validators you run, not the amount of coins you have on each validator.
- If you have more than 32 ETH you can get slashed more than 32 ETH. So no reason to stake more than 32 ETH per validator.
- Essentially, you just want to run as many validators as you can in lumps of 32 if you are trying to maximise yield per ETH staked.
- Staking won’t happen unless more than 100k ETH are staked (threshold).
Validator setup
- 3 components of a basic validator setup
- Beacon Node = Stores Chain State (think Eth full nodes).
- Validator = Node registered with Eth 2.0. Doesn’t need to store the chain. Lightweight, need to trust a Beacon Node.
- Signer = Connects to or is part of a Validator
Implications
- Validators have are disincentivised to stake more than the minimum of 32 ETH (can be slashed for full amount, so minimise the amount at risk).
- Cheating or unintentional cheating can be very expensive. Eg accidentally double signing (running hot hot and one goes down etc).
- So, to get the most juice, run more validators.
- A fund holding 100k ETH = Need to run 3,125 validators! Significant burden on infra. With sharding, this could get hairy – up to 64x particularly on Beacon nodes.
- Infra will be expensive.
- Rewards: Up to 11% annual staking returns (if staking 32 ETH on a validator. If stake eg 64 ETH on one validator, then rewards would be max 5.5%)
What’s next?
- First multi-client testnet of Eth 2.0
- Beacon Chain rolling out as soon as July
- Start preparing early!
Alchemy
- Alchemy’s goal will eventually be to make it easy for retail investors who have at least 32 ETH to easily become a validator.
Cloud costs
- ETH full node might cost $500/mth on cloud providers, most of that is on storage.
- Validator (not beacon node) estimated at sub $100 monthly costs.
My thoughts (not part of the presentation)
- To be a profitable validator, your revenue needs to exceed your costs.
- Your annual revenue per validator is (based on 11% interest) 11% x 32 ETH = 3.52 ETH.
- So your profit is 3.52 ETH less costs of running a validator (cloud / hardware costs, management time etc)
- Let’s say your cost is $100 / mth, = $1,200 per year (mentioned in the presentation)
- To break even, 3.52 ETH needs to equal $1,200
- This implies a break-even price of $1,200/3.52 ETH = $340 per ETH
- Plus you need to account for the risk of getting slashed.
- If ETH is below $340, it’s not financially worth you staking.
- Obviously this number is sensitive to your costs of running a validator node. If your costs are half that (eg $50/month) then you will be profitable at $170/ETH
- There is a price of ETH below which it becomes loss-making to validate (just like Bitcoin).
- The real winners are the cloud providers!