Robust remote data checking

Abstract

Remote data checking protocols, such as provable data possession (PDP) [1], allow clients that outsource data to untrusted servers to verify that the server continues to correctly store the data. Through the careful integration of forward error-correcting codes and remote data checking, a system can prove possession with arbitrarily high probability. We formalize this notion in the robust data possession guarantee. We distill the key performance and security requirements for integrating forward error-correcting codes into PDP and describe an encoding scheme and file organization for robust data possession that meets these requirements. We give a detailed analysis of this scheme and build a Monte-Carlo simulation to evaluate tradeoffs in reliability, space overhead, and performance. A practical way to evaluate these tradeoffs is an essential input to system design, allowing the designer to choose the encoding and data checking protocol parameters that realize robust data possession.