Privacy-Preserving Coded Mobile Edge Computing for Low-Latency Distributed Inference

Abstract

We consider a mobile edge computing scenario where a number of devices want to perform a linear inference <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">${W}{x} $ </tex-math></inline-formula> on some local data <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$ {x}$ </tex-math></inline-formula> given a network-side matrix <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$ {W}$ </tex-math></inline-formula> . The computation is performed at the network edge over a number of edge servers. We propose a coding scheme that provides information-theoretic privacy against <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$z$ </tex-math></inline-formula> colluding (honest-but-curious) edge servers, while minimizing the overall latency—comprising upload, computation, download, and decoding latency—in the presence of straggling servers. The proposed scheme exploits Shamir’s secret sharing to yield data privacy and straggler mitigation, combined with replication to provide spatial diversity for the download. We also propose two variants of the scheme that further reduce latency. For a considered scenario with 9 edge servers, the proposed scheme reduces the latency by 8% compared to the nonprivate scheme recently introduced by Zhang and Simeone, while providing privacy against an honest-but-curious edge server.