Specification and Verification of Side Channel Declassification

TL;DR

This paper introduces a semantic model and verification techniques for programs that intentionally leak information via side channels, balancing security and efficiency through declassification policies and program structuring.

Contribution

It proposes a formal model for side-channel declassification, enabling verification with existing tools, and introduces a program 'manifest form' to simplify security verification.

Findings

Semantic model for side-channel declassification

Verification of declassification policies using model checking

Program structuring as a tradeoff for easier verification

Abstract

Side channel attacks have emerged as a serious threat to the security of both networked and embedded systems -- in particular through the implementations of cryptographic operations. Side channels can be difficult to model formally, but with careful coding and program transformation techniques it may be possible to verify security in the presence of specific side-channel attacks. But what if a program intentionally makes a tradeoff between security and efficiency and leaks some information through a side channel? In this paper we study such tradeoffs using ideas from recent research on declassification. We present a semantic model of security for programs which allow for declassification through side channels, and show how side-channel declassification can be verified using off-the-shelf software model checking tools. Finally, to make it simpler for verifiers to check that a program conforms to a particular side-channel declassification policy we introduce a further tradeoff between efficiency and verifiability: by writing programs in a particular "manifest form" security becomes considerably easier to verify.