Statistical Profiling of Micro-Architectural Traces and Machine Learning for Spectre Detection: A Systematic Evaluation

Abstract

Spectre attacks exploit features of modern processors to leak sensitive data through speculative execution and shared resources (such as caches). A popular approach to detect such attacks deploys Machine Learning (ML) to identify suspicious micro-architectural patterns. These techniques, however, are often rather ad-hoc in terms of the selection of micro-architectural features as well as ML techniques and frequently lack a description of the underlying training- and test-data. To address these shortcomings, we systematically evaluate a large range of (combinations of) micro-architectural features recorded in up to 40 Hardware Performance Counters (HPCs), as well as multiple ML algorithms on a comprehensive set of scenarios and datasets. Using statistical methods, we rank the HPCs used to generate our dataset, which helps us determine the minimum number of features required for detecting Spectre attacks with high accuracy and minimal overhead. Furthermore, we identify the best-performing ML classifiers, and provide a comprehensive description of our data collection, running scenarios, selected HPCs, and chosen classification models.