Biblio
Memory-safety violations are the primary cause of security and reliability issues in software systems written in unsafe languages. Given the limited adoption of decades-long research in software-based memory safety approaches, as an alternative, Intel released Memory Protection Extensions (MPX)–-a hardware-assisted technique to achieve memory safety. In this work, we perform an exhaustive study of Intel MPX architecture along three dimensions: (a) performance overheads, (b) security guarantees, and (c) usability issues. We present the first detailed root cause analysis of problems in the Intel MPX architecture through a cross-layer dissection of the entire system stack, involving the hardware, operating system, compilers, and applications. To put our findings into perspective, we also present an in-depth comparison of Intel MPX with three prominent types of software-based memory safety approaches. Lastly, based on our investigation, we propose directions for potential changes to the Intel MPX architecture to aid the design space exploration of future hardware extensions for memory safety. A complete version of this work appears in the 2018 proceedings of the ACM on Measurement and Analysis of Computing Systems.
Content-based routing (CBR) is a powerful model that supports scalable asynchronous communication among large sets of geographically distributed nodes. Yet, preserving privacy represents a major limitation for the wide adoption of CBR, notably when the routers are located in public clouds. Indeed, a CBR router must see the content of the messages sent by data producers, as well as the filters (or subscriptions) registered by data consumers. This represents a major deterrent for companies for which data is a key asset, as for instance in the case of financial markets or to conduct sensitive business-to-business transactions. While there exists some techniques for privacy-preserving computation, they are either prohibitively slow or too limited to be usable in real systems. In this paper, we follow a different strategy by taking advantage of trusted hardware extensions that have just been introduced in off-the-shelf processors and provide a trusted execution environment. We exploit Intel's new software guard extensions (SGX) to implement a CBR engine in a secure enclave. Thanks to the hardware-based trusted execution environment (TEE), the compute-intensive CBR operations can operate on decrypted data shielded by the enclave and leverage efficient matching algorithms. Extensive experimental evaluation shows that SGX adds only limited overhead to insecure plaintext matching outside secure enclaves while providing much better performance and more powerful filtering capabilities than alternative software-only solutions. To the best of our knowledge, this work is the first to demonstrate the practical benefits of SGX for privacy-preserving CBR.