[Paper Review] ScaRR: Scalable Runtime Remote Attestation for Complex Systems

Flavio Toffalini, Eleonora Losiouk+3 more|arXiv (Cornell University)|Jul 20, 2018

Security and Verification in Computing52 references4 citations

TL;DR

ScaRR is the first scalable runtime remote attestation scheme for complex systems, such as cloud-based virtual machines, that enables fine-grained, intermediate verification of control-flow integrity through a novel fragmentation-based control-flow model. It achieves an average of 2 million control-flow events per second, significantly outperforming prior solutions while supporting multi-threading and dynamic runtime analysis.

ABSTRACT

The introduction of remote attestation (RA) schemes has allowed academia and industry to enhance the security of their systems. The commercial products currently available enable only the validation of static properties, such as applications fingerprint, and do not handle runtime properties, such as control-flow correctness. This limitation pushed researchers towards the identification of new approaches, called runtime RA. However, those mainly work on embedded devices, which share very few common features with complex systems, such as virtual machines in a cloud. A naive deployment of runtime RA schemes for embedded devices on complex systems faces scalability problems, such as the representation of complex control-flows or slow verification phase. In this work, we present ScaRR: the first Scalable Runtime Remote attestation schema for complex systems. Thanks to its novel control-flow model, ScaRR enables the deployment of runtime RA on any application regardless of its complexity, by also achieving good performance. We implemented ScaRR and tested it on the benchmark suite SPEC CPU 2017. We show that ScaRR can validate on average 2M control-flow events per second, definitely outperforming existing solutions.

Motivation & Objective

• To address the lack of scalable runtime remote attestation (RA) solutions for complex systems like cloud VMs, which are not adequately supported by existing embedded-device-focused RA schemes.
• To enable runtime verification of control-flow integrity without interrupting program execution, supporting dynamic and intermediate validation.
• To overcome scalability limitations of prior RA schemes that rely on monolithic path hashing or symbolic execution, which are infeasible for complex software.
• To support fine-grained attack detection by enabling the Verifier to analyze partial execution reports with minimal latency.
• To design a solution compatible with modern system features such as multi-threading and standard security mechanisms like ASLR and stack canaries.

Proposed method

• ScaRR introduces a novel control-flow model that fragments the complete execution path into meaningful sub-paths, enabling partial reporting.
• The Prover generates and sends intermediate partial reports at runtime, each representing a segment of the control-flow execution.
• The Verifier performs lightweight, sound verification of each partial report using a combination of static analysis and shadow execution to ensure correctness.
• The system leverages compiler instrumentation via LLVM to insert monitoring code into the application source, enabling runtime control-flow event collection.
• ScaRR uses a trusted anchor to ensure report authenticity and integrity, while relying on standard OS-level protections (e.g., W^X, ASLR) for runtime defense.
• The architecture supports multi-threaded execution by independently tracking and reporting control-flow events per thread.

Experimental results

Research questions

• RQ1 Can a runtime remote attestation scheme be designed to scale effectively for complex systems such as cloud-based virtual machines?
• RQ2 How can control-flow integrity be verified in real time without relying on monolithic path hashing or expensive symbolic execution?
• RQ3 Can intermediate verification of execution paths be achieved without disrupting program execution, enabling fine-grained attack detection?
• RQ4 What is the performance ceiling of a runtime RA scheme in terms of attestation and verification speed for complex workloads?
• RQ5 How does the proposed control-flow representation model compare in security and scalability to existing approaches like C-FLAT or DIAT?

Key findings

• ScaRR achieves an average attestation speed of 2 million control-flow events per second, significantly outperforming prior solutions that manage only a few hundred to a few thousand per second.
• The verification phase is highly efficient due to lightweight checks on partial reports, enabling real-time validation without performance bottlenecks.
• ScaRR generates limited network traffic, as partial reports are compact and only a subset of control-flow events are transmitted per interval.
• The control-flow model based on path fragmentation enables accurate representation of complex control-flow graphs without exponential growth in representation size.
• ScaRR supports multi-threaded applications and maintains security guarantees even under a fully compromised user-space threat model.
• The approach is resilient to attacks like COOP, which exploit loss of execution order in multi-set based models, due to its sound verification via shadow execution.