Architecture Support for FPGA Multi-tenancy in the Cloud

TL;DR

This paper introduces an FPGA architecture supporting multi-tenancy in cloud environments, enabling efficient space-sharing among tenants with high throughput and increased utilization, while maintaining performance.

Contribution

It proposes a novel NoC-based FPGA architecture for multi-tenancy, demonstrating high performance and significantly improved FPGA utilization in cloud deployments.

Findings

Achieved 2x higher maximum frequency than state-of-the-art

Demonstrated 25.6 Gbps bandwidth for data movement

Realized 6x higher FPGA utilization in case study

Abstract

Cloud deployments now increasingly provision FPGA accelerators as part of virtual instances. While FPGAs are still essentially single-tenant, the growing demand for hardware acceleration will inevitably lead to the need for methods and architectures supporting FPGA multi-tenancy. In this paper, we propose an architecture supporting space-sharing of FPGA devices among multiple tenants in the cloud. The proposed architecture implements a network-on-chip (NoC) designed for fast data movement and low hardware footprint. Prototyping the proposed architecture on a Xilinx Virtex Ultrascale+ demonstrated near specification maximum frequency for on-chip data movement and high throughput in virtual instance access to hardware accelerators. We demonstrate similar performance compared to single-tenant deployment while increasing FPGA utilization ( we achieved 6x higher FPGA utilization with our case study), which is one of the major goals of virtualization. Overall, our NoC interconnect achieved about 2x higher maximum frequency than the state-of-the-art and a bandwidth of 25.6 Gbps.