A lightweight MapReduce framework for secure processing with SGX

摘要

MapReduce is a programming model used extensively for parallel dataprocessing in distributed environments. A wide range of algorithms wereimplemented using MapReduce, from simple tasks like sorting and searching up tocomplex clustering and machine learning operations. Many of theseimplementations are part of services externalized to cloud infrastructures.Over the past years, however, many concerns have been raised regarding thesecurity guarantees offered in such environments. Some solutions relying oncryptography were proposed for countering threats but these typically imply ahigh computational overhead. Intel, the largest manufacturer of commodity CPUs,recently introduced SGX (software guard extensions), a set of hardwareinstructions that support execution of code in an isolated secure environment.In this paper, we explore the use of Intel SGX for providing privacy guaranteesfor MapReduce operations, and based on our evaluation we conclude that itrepresents a viable alternative to a cryptographic mechanism. We presentresults based on the widely used k-means clustering algorithm, but ourimplementation can be generalized to other applications that can be expressedusing MapReduce model.