Tommaso Frassetto’s research while affiliated with Technical University of Darmstadt and other places

Smart contracts enable users to execute payments depending on complex program logic. Ethereum is the most notable example of a blockchain that supports smart contracts leveraged for countless applications including games, auctions and financial products. Unfortunately, the traditional method of running contract code on-chain is very expensive, for instance, on the Ethereum platform, fees have dramatically increased, rendering the system unsuitable for complex applications. A prominent solution to address this problem is to execute code off-chain and only use the blockchain as a trust anchor. While there has been significant progress in developing off-chain systems over the last years, current off-chain solutions suffer from various drawbacks including costly blockchain interactions, lack of data privacy, huge capital costs from locked collateral, or supporting only a restricted set of applications. In this paper, we present POSE -- a practical off-chain protocol for smart contracts that addresses the aforementioned shortcomings of existing solutions. POSE leverages a pool of Trusted Execution Environments (TEEs) to execute the computation efficiently and to swiftly recover from accidental or malicious failures. We show that POSE provides strong security guarantees even if a large subset of parties is corrupted. We evaluate our proof-of-concept implementation with respect to its efficiency and effectiveness.

Performing machine learning tasks in mobile applications yields a challenging conflict of interest: highly sensitive client information (e.g., speech data) should remain private while also the intellectual property of service providers (e.g., model parameters) must be protected. Cryptographic techniques offer secure solutions for this, but have an unacceptable overhead and moreover require frequent network interaction. In this work, we design a practically efficient hardware-based solution. Specifically, we build Offline Model Guard (OMG) to enable privacy-preserving machine learning on the predominant mobile computing platform ARM - even in offline scenarios. By leveraging a trusted execution environment for strict hardware-enforced isolation from other system components, OMG guarantees privacy of client data, secrecy of provided models, and integrity of processing algorithms. Our prototype implementation on an ARM HiKey 960 development board performs privacy-preserving keyword recognition using TensorFlow Lite for Microcontrollers in real time.

The increasing complexity of modern computing devices has rendered security architectures vulnerable to recent side-channel and transient-execution attacks. We discuss the most relevant defenses as well as their drawbacks and how to overcome them for next-generation secure processor design.

Fault-injection attacks have been proven in the past to be a reliable way of bypassing hardware-based security measures, such as cryptographic hashes, privilege and access permission enforcement, and trusted execution environments. However, traditional fault-injection attacks require physical presence, and hence, were often considered out of scope in many real-world adversary settings. In this paper we show this assumption may no longer be justified. We present V0LTpwn, a novel hardware-oriented but software-controlled attack that affects the integrity of computation in virtually any execution mode on modern x86 processors. To the best of our knowledge, this represents the first attack on x86 integrity from software. The key idea behind our attack is to undervolt a physical core to force non-recoverable hardware faults. Under a V0LTpwn attack, CPU instructions will continue to execute with erroneous results and without crashes, allowing for exploitation. In contrast to recently presented side-channel attacks that leverage vulnerable speculative execution, V0LTpwn is not limited to information disclosure, but allows adversaries to affect execution, and hence, effectively breaks the integrity goals of modern x86 platforms. In our detailed evaluation we successfully launch software-based attacks against Intel SGX enclaves from a privileged process to demonstrate that a V0LTpwn attack can successfully change the results of computations within enclave execution across multiple CPU revisions.

Recent research has demonstrated that Intel's SGX is vulnerable to software-based side-channel attacks. In a common attack, the adversary monitors CPU caches to infer secret-dependent data accesses patterns. Known defenses have major limitations, as they require either error-prone developer assistance, incur extremely high runtime overhead, or prevent only specific attacks. In this paper, we propose data location randomization as a novel defense against side-channel attacks that target data access patterns. Our goal is to break the link between the memory observations by the adversary and the actual data accesses by the victim. We design and implement a compiler-based tool called DR.SGX that instruments the enclave code, permuting data locations at fine granularity. To prevent correlation of repeated memory accesses we periodically re-randomize all enclave data. Our solution requires no developer assistance and strikes the balance between side-channel protection and performance based on an adjustable security parameter.