Plasma chain for offchain gas payments


#1

TLDR: We suggest using a plasma chain to pay for gas offchain.

Construction

Let P be a plasma chain which accepts token deposits (e.g. ETH or ERC20s) to top-up “gas balances” for offchain gas payments. A user that wants to pay for the gas of some transaction T using his gas balance on P creates a plasma “gas transaction” containing:

  1. Target transaction: The transaction T, which can have a “native” gas price of 0 Gwei.
  2. Gas account: The account from which the offchain gas payment will be deducted.
  3. Meta gas price: The offered “meta” gas price for the miner that includes T onchain.
  4. Timeout: An optional timeout (a block height) after which the offer ends.
  5. Signature: A signature of the gas transaction against the gas account.

Once the gas transaction has been included in P the miner that includes T onchain is eligible to claim the corresponding gas payment. For this, the miner creates a “gas receipt transaction” on P proving that T was confirmed onchain by him and deducting the appropriate balance.

Specifically, the plasma chain has light-client access to the main chain and the miner provides the Merkle path in the receipt trie corresponding to T and uses gasUsed to calculate the meta gas payment. Notice that the nonce of T means that there can only be a single miner claiming the meta gas payment.

Discussion

Offchain gas payments can been useful for a variety of reasons. First of all they expand the design space of gas payments, allowing to pay gas with non-ETH tokens and to create exotic conditions for gas payment such as the optional timeout in the construction above.

Second, offchain gas payments can be useful for cheaply paying for gas without touching the EVM’s state trie in the context of state-minimised execution engines or stateless blob shards.

It is also possible to setup offchain gas payments with payment channels to individual miners but that encourages centralisation. The plasma construction is a decentralisation-friendly approach to offchain gas payments.


#2

I like this. It works particularly well for slow transactions.

Potential problems basically boil down to usability. Users have to wait for Ethereum finality before their request tx is included in the child chain. As a result, any transactions made through the system will be necessarily slow.

Root chain ETH is probably preferable to child chain ETH (locked up for at least challenge period, need to manage the whole child chain), so my guess is that the biggest advantage here is the ability to pay for gas with value other than ETH. However, it might be hard to convince pools to accept non-ETH fees. Small pool miners probably won’t want to deal with receiving their mining rewards in a bunch of different currencies. You could fix this by allowing users to specify what coins they’re ok with receiving, but that places some significant complexity on the pool admin. Currently >90% of blocks are mined by pools, so I’d be interested to hear the pool admin perspective.

The challenge period also complicates things. Miners might not want to wait for the exit to finalize to receive their rewards. An atomic swap of Plasma coins for on-chain coins (for a fee) could fix this issue. (I’m actually not really sure if this is a problem - might be something to discuss with a large-scale miner).

My final concern depends on the consensus mechanism. If we’re assuming a PoA-operator system, then it’s possible for the operator to withhold a fee from a miner by refusing to include their gas receipt transaction. Fortunately, this probably just results in the miner losing fees for a single block and then refusing to mine any more transactions via that gas chain.