A Smart Contract Refereed Data Retrieval Protocol with a Provably Low Collateral Requirement

James Shook, Scott Simon, and Peter Mell

Abstract

We present a protocol for a cryptoeconomic fair exchange of data previously owned by the purchaser for tokens that functions even when both parties are anonymous. This enables peer-to-peer data storage without identity verification. We use a smart contract on a decentralized ledger as a trusted third party. Actual data transfer can take place with any standard anonymous exchange channel. Due to the anonymity of the parties, the smart contract cannot punish either party's off-ledger reputation. Furthermore, the contract has limited power to arbitrate fault in off-ledger disputes. Thus, an important feature of our protocol is a collateral mechanism that collectively punishes both Alice and Bob if either of them abandons the protocol or cheats. However, we prove that parameters can be chosen such that the collateral can be made, subject to data size, arbitrarily low and still result in an expected financial loss if either Alice or Bob cheats. We are able to achieve this due to our non-standard use of error-correcting codes. In addition, the protocol allows those storing the data to exchange it without the client's participation.

Available format(s)
Category
Applications
Publication info
Preprint. Minor revision.
Keywords
blockchaincollateraldecentralized ledgerdecentralized storageerror-correcting codesfair exchangesmart contract
Contact author(s)
james shook @ nist gov
sbsimon @ tycho ncsc mil
History
Short URL
https://ia.cr/2019/541

CC BY

BibTeX

@misc{cryptoeprint:2019/541,
author = {James Shook and Scott Simon and Peter Mell},
title = {A Smart Contract Refereed Data Retrieval Protocol with a Provably Low Collateral Requirement},
howpublished = {Cryptology ePrint Archive, Paper 2019/541},
year = {2019},
note = {\url{https://eprint.iacr.org/2019/541}},
url = {https://eprint.iacr.org/2019/541}
}

Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.