作者: Craig Costello , Cedric Fournet , Jon Howell , Markulf Kohlweiss , Benjamin Kreuter
DOI: 10.1109/SP.2015.23
关键词: Cryptography 、 Parallel computing 、 Mathematical proof 、 Computation 、 Compiler 、 Verifiable computing 、 Computer science 、 Cryptographic primitive 、 Bounded function 、 Correctness 、 Cryptographic protocol 、 Cloud computing
摘要: Cloud computing sparked interest in Verifiable Computation protocols, which allow a weak client to securely outsource computations remote parties. Recent work has dramatically reduced the client's cost verify correctness of their results, but overhead produce proofs remains largely impractical. Geppetto introduces complementary techniques for reducing prover and increasing flexibility. With Multi QAPs, reduces sharing state between (e.g., For MapReduce) or within single computation by up two orders magnitude. Via careful choice cryptographic primitives, Geppetto's instantiation bounded proof bootstrapping improves on prior bootstrapped systems five magnitude, albeit at some universality. also efficiently verifies correct execution proprietary (i.e., Secret) algorithms. Finally, use energy-saving circuits brings prover's costs more line with program's actual (rather than worst-case) time. is implemented full-fledged, scalable compiler runtime that consume LLVM code generated from variety source C programs libraries.