A Memory Encryption Engine Suitable for General Purpose Processors

By Shay Gueron. 2016

232 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Ekiden: A Platform for Confidentiality-Preserving, Trustworthy, and Performant Smart Contracts (2018, neutral): “The SGX memory encryption engine also guarantees data integrity and prevents memory replay attacks [39].”
Keystone: an open framework for architecting trusted execution environments (2020, methodology): “The protection can be done either in the software as a part of the Keystone RT (Figure 5(d)) or by a dedicated trusted hardware unit—a memory encryption engine (MEE) [44]—with the SM’s on-chip memory capability (Figure 5).”
Eleos: ExitLess OS Services for SGX Enclaves (2017, background): “We speculate that this is due to the SGX memory encryption engine storing its integrity tree in the EPC [15], which competes for LLC space as well.”
Plundervolt: Software-based Fault Injection Attacks against Intel SGX (2020, background): “This has reached a point where Intel ultimately considers main memory as an untrusted storage facility and fully encrypts and authenticates all memory within the Intel SGX enclave security architecture [24].”
Intel SGX Explained (2016, background): “This paper does not reflect the information available in two papers [74, 109] that were published after the first version of this paper.”
YerbaBuena: Securing Deep Learning Inference Data via Enclave-based Ternary Model Partitioning (2018, unknown): “We can also find the technical details about the SGX attestation and sealing mechanisms in [1], dynamic memory allocation of SGX2 in [34], andMemory Encryption Engine in [15].”
Occlumency: Privacy-preserving Remote Deep-learning Inference Using SGX (2019, methodology): “The size of PRM is limited to 128 MB, among which the hardware carves out ≈ 90 MB as the Enclave Page Cache (EPC) and the rest is reserved for the meta-data required by the Memory Encryption Engine (MEE) [32].”
Hacking in Darkness: Return-oriented Programming against Secure Enclaves (2017, background): “In addition, a memory encryption engine (MEE) [11, 15] that is a extension of the memory controller encrypts enclave code and data before they are being written to main memory.”
Foreshadow: Extracting the Keys to the Intel SGX Kingdom with Transient Out-of-Order Execution (2018, unknown): “As such, the MEE hardware unit does not impose any fundamental limitations on the Foreshadow attack, and is assuredly not the cause for the observation that we cannot read enclave secrets residing in the L2 cache.”
Another Flip in the Wall of Rowhammer Defenses (2017, methodology): “The encryption by the Memory Encryption Engine (MEE) is transparent to the processor’s cores [26].”