Provably Secure Computation via Isolated Execution Environments [20190318]
iSpeech.org
[CENTRA Webinar] "Provably Secure Computation via Isolated Execution Environments" was held on March 18th, 2019 at 9-10pm US EDT by Global *CENTRA project, via ACIS Lab, University of Florida. *http://globalcentra.org
[Abstract]
New capabilities of modern hardware technologies allow for the execution of arbitrary code within environments isolated from the rest of the system, and provide cryptographic mechanisms for reporting on these executions to remote parties. The rigorous security analysis of these protocols requires the development of scalable proof techniques for dealing with new forms of composition of cryptographic primitives arising in this setting.
The attestation guarantees of trusted hardware technologies can be used to bootstrap secure communications between participants and the hardware trust anchor. With a small and functionality-agnostic bootstrapping procedure, the bulk of the computational load can be moved to an untrusted party equipped with trusted hardware, an attractive feature for Cloud-based scenarios.
This presentation describes the first contributions in this direction, formalising the intuition of using hardware-based isolation and attestation guarantees to construct a new generation of efficient cryptographic protocols for searchable encryption, secure outsourcing of computation and secure multiparty computation. We will also overview secure hardware solutions available on the market and discuss the extent to which they match our formal model of security for trusted hardware. Open challenges include designing solutions relying upon weaker hardware assumptions, and developing a heterogeneous, scalable and fault-tolerant attestation-based system.
[Speaker]
Bernado Portela is a post-doctoral researcher at HASLab/INESC TEC (https://haslab.uminho.pt ), Portugal. He holds a M.Sc. thesis in Informatics Engineering from University of Minho, and a Ph.D. under the MAPi doctoral programme.
Dr. Portela's research interests are cryptography and information security, more specifically regarding secure multiparty computation protocols and trusted hardware. My goals are to improve the implementation of high-assurance multiparty computation protocols, by bridging the gap between rigorous theoretical security models and efficient practical implementations. Furthermore, I am interested in the exploring the topic of secure machine learning, both from the scope of privacy-preserving computation and from the scope of differential privacy.
Relevant contributions in this context include the first provable security approach for formalizing security guarantees of Isolated Execution Environments (IEEs), and the first general approach to implementing MPC protocols using IEE-enabled systems.
source
Gloss