Scholarlink

<p>Field-programmable gate arrays (FPGAs) are increasingly utilized in cloud computing due to their reconfigurability and ability to accelerate tasks such as machine learning and cryptography. However, multi-tenancy, where multiple users share FPGA resources, introduces significant security challenges. This paper analyzes both traditional and FPGA-specific attacks, including side-channel, covert-channel, and fault injection attacks, which exploit shared resources to compromise confidentiality, integrity, and availability. Countermeasures, encompassing preventive and runtime approaches, are reviewed to address these threats, emphasizing the need for a multi-layered security strategy. Although multi-tenancy in FPGA cloud servers has yet to be widely adopted, this paper provides critical insights into securing future multi-tenant FPGA infrastructures, ensuring the safe and efficient use of FPGA resources in cloud environments.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Toward secure program execution in multi-tenant cloud FPGA environments

  • 期刊论文

  • 发表日期：

    2025年5月21日

查看全文
  • Tian Chen,
  • Yu-an Tan,
  • Wenjuan Li,
  • Zhaohui Ci,
  • Ning Shi

摘要

Field-programmable gate arrays (FPGAs) are increasingly utilized in cloud computing due to their reconfigurability and ability to accelerate tasks such as machine learning and cryptography. However, multi-tenancy, where multiple users share FPGA resources, introduces significant security challenges. This paper analyzes both traditional and FPGA-specific attacks, including side-channel, covert-channel, and fault injection attacks, which exploit shared resources to compromise confidentiality, integrity, and availability. Countermeasures, encompassing preventive and runtime approaches, are reviewed to address these threats, emphasizing the need for a multi-layered security strategy. Although multi-tenancy in FPGA cloud servers has yet to be widely adopted, this paper provides critical insights into securing future multi-tenant FPGA infrastructures, ensuring the safe and efficient use of FPGA resources in cloud environments.