SE-PIM: In-Memory Acceleration of Data-Intensive Confidential Computing

Demand for data-intensive workloads and confidential computing are the prominent research directions shaping future of cloud computing. Computer architectures evolving to accommodate large data better. Protecting computation sensitive is also an imperative yet challenging objective; processor-supported secure enclaves serve as key element in cloud. However, side-channel attacks threatening their security boundaries. The current processor consume a considerable portion its cycles moving data. Near promising approach that minimizes redundant movement by placing inside storage. In this paper, we present novel design Processing-In-Memory (PIM) workload accelerator Based on our observation closer memory can achieve efficiency confidentiality processed information simultaneously, study advantages \emph{inside} memory. We then explain model programming developed PIM-based offloading. construct findings into software-hardware co-design, which call PIM-Enclave. Our illustrates acceleration. evaluation shows PIM-Enclave provide resistant offloading run applications with negligible performance overhead compared baseline PIM model.