dogethereum / dogethereum-agents Goto Github PK

Hello great work on this. I think fraud proofs similar to a challenge response / plasma with bonds is a way to ensure proposed headers or groups of headers follows the probable longest chain. The other way is a BFT on the proposed headers similar to how systems such as komodo do to notarize sidechain hashes to push up to kmd and then bitcoin. The problem with that approach is that every notary needs to know about all sidechains and a cryptoeconomic system is a bit more flexible instead. Regardless on this system we dont really have problems with managing 2 chains unless we start to get to a situation where we want to abstract it to manage any number of sidechains with adaptors to different consensus algorithms and codebases. I think a few projects are also doing this.

When I think about SPV client security unfortunately there hasnt been many proper research done here with published papers on this subject. As far as I can see I cannot really get around a fundamental blocking problem with SPV in that a bad miner can create a proof that follows difficulty rules and builds a chain on (I make a bad tx then build 6 blocks on it on a fork or a private chain then push to SPV clients). since this is an autonomous system it exasberates the problem and incentives for bad actors.

To my question now.. in this system what stops either a challenger which is bad and decides to block good blocks while mining bad blocks and submitting? The pow would be satisfied and the attack would involve just creating a tx that spends or double spends doge to eth as eth inflates free money. afaics the eth chain wont know which one to choose the bad block vs good block because it satisfies the target pow and merkle roots would be correct both building on a previous superblock.

In the case of a bad actor how could one create a challenge in a way that the system would differentiate a good vs bad block? In my mind the only way is the number of challenger good blocks being the same must > submitter good blocks which resolves to a BFT effectively.

Imagine I create a bad sequence of superblocks in an attempt to inflate dogeeth by 1 million. Now a challenger will come around and challenge me on the first superblock that doesnt match.. how will it know which actor is bad? It follows the protocol rules just like the good challenger. Without the utxo db you dont have a source for global truth to resolve?

Some benefits:

Dynamic gas price providers
Review FederatorSupport.updateContractFacadesGasPrice()
Multi-dimensional arrays
maybe refactor bulkStoreHeaders()
hyperledger/web3j#404
Adds web3j shutdown method hyperledger/web3j#433
Expose additional behaviors from auto-generated contract classes for use with ethGetLogs hyperledger/web3j#387

Currently, they send a deposit before doing each action.
Create a mechanism where they can deposit and withdraw eth once a day or something like that.