Yukui Luo’s PhD Proposal Review

Abstract:

With the widespread adoption of cloud computing, the demand for programmable hardware acceleration devices, such as field-programmable gate array (FPGA), has increased. To further improve the performance of FPGA-enabled cloud computing, one promising technology is to virtualize the hardware resources of an FPGA device, which allows multiple users to share the same FPGA. This solution can provide on-demand instances at the FPGA resource and time levels, significantly improving the utilization and energy efficiency of the FPGA devices. However, due to the hardware reconfigurability of FPGA, current virtualization methods for multi-tenant GPU and TPU instances are incompatible with multi-tenant FPGA virtualization.
We define the threat model for multi-tenant FPGA and discuss the security issues related to Confidentiality, Data Integrity, and Availability. Based on an analysis of potential attacks, we present our latest research results and propose two future research directions for mitigations: (1) a multi-tenant FPGA plug-to-play obfuscation module and (2) a hardware-software co-designed multi-tenant FPGA virtualization system, which includes a hypervisor and a smart multi-tenant FPGA platform.