Portal de Periódicos da CAPES

Reparo: Publicly Verifiable Layer to Repair Blockchains

2021; Springer Science+Business Media; Linguagem: Inglês

10.1007/978-3-662-64331-0_2

1611-3349

Sri Aravinda Krishnan Thyagarajan, Adithya Bhat, Bernardo Magri, Daniel Tschudi, Aniket Kate,

Cloud Data Security Solutions

Although blockchains aim for immutability as their core feature, several instances have exposed the harms with perfect immutability. The permanence of illicit content inserted in Bitcoin poses a challenge to law enforcement agencies like Interpol, and millions of dollars were lost in buggy smart contracts in Ethereum. A line of research then spawned on redactable blockchains with the aim of solving the problem of redacting illicit contents from both permissioned and permissionless blockchains. However, all the existing proposals follow the build-new-chain approach for redactions, and cannot be integrated with existing running blockchains, such as Bitcoin and Ethereum. This work demonstrates that the traditional build-new-chain approach for blockchain redactions is not necessary. We present $$\mathsf {Reparo}$$ (In the Harry Potter universe, ‘ $$\mathsf {Reparo}$$ ’ is a spell that repairs objects), a publicly verifiable layer on top of any blockchain to perform repairs, ranging from fixing buggy contracts to removing illicit contents from the chain. We present an efficient instantiation of $$\mathsf {Reparo}$$ over Ethereum (with proof of work based consensus) for repairing smart contract bugs. In this protocol, any Ethereum user may propose a repair and a deliberation process ensues resulting in a decision that complies with the repair policy of the chain and is publicly verifiable. A repair operation (for instance, fixing a bug in a contract) is then performed according to the repair proposal and the state of Ethereum is updated accordingly. $$\mathsf {Reparo}$$ ’s advantages are multi-fold: (i) Since $$\mathsf {Reparo}$$ follows a layer design, it helps facilitate additional functionalities for Ethereum while maintaining the same provable security guarantees; (ii) $$\mathsf {Reparo}$$ can be easily tailored to different consensus requirements (like proof of stake), does not require heavy cryptographic machinery, and thus, can be integrated with other existing blockchains (such as Bitcoin, Cardano) as well. We evaluate $$\mathsf {Reparo}$$ with Ethereum mainnet and show that the cost of fixing several prominent smart contract bugs is almost negligible. For instance, the cost of repairing the prominent Parity Multisig wallet bug with $$\mathsf {Reparo}$$ is as low as $$0.00005\%$$ of the Ethers that can be retrieved after the fix.