Remote Attestation with Domain-Based Integrity Model and Policy Analysis

We propose and implement an innovative remote attestation framework called DR@FT for efficiently measuring a target system based on an information flow-based integrity model. With this model, the high integrity processes of a system are first measured and verified, and these processes are then protected from accesses initiated by low integrity processes. Toward dynamic systems with frequently changed system states, our framework verifies the latest state changes of a target system instead of considering the entire system information. Our attestation evaluation adopts a graph-based method to represent integrity violations, and the graph-based policy analysis is further augmented with a ranked violation graph to support high semantic reasoning of attestation results. As a result, DR@FT provides efficient and effective attestation of a system's integrity status, and offers intuitive reasoning of attestation results for security administrators. Our experimental results demonstrate the feasibility and practicality of DR@FT.

1 Introduction

In distributed computing environments, it is crucial to measure whether a remote party runs buggy, malicious application codes or is improperly configured. Remote attestation techniques have been proposed for this purpose by analyzing the integrity of a remote system to determine its trustworthiness. Typical attestation mechanisms are designed based on the following steps. First, an attestation requester (attester) sends a challenge to a target system (attestee), which responds with the evidence of the integrity of its hardware and software components. Second, the attester derives runtime properties of the attestee and determines the trustworthiness of the attestee. Finally and optionally, the attester returns an attestation result, such as integrity status, to the attestee.