3 papers
cs.AR2024
Arcus: SLO Management for Accelerators in the Cloud with Traffic Shaping
Jiechen Zhao, Ran Shu, Katie Lim +4
Cloud servers use accelerators for common tasks (e.g., encryption, compression, hashing) to improve CPU/GPU efficiency and overall performance. However, users' Service-level Object…
cs.AR2024
Beehive: A Flexible Network Stack for Direct-Attached Accelerators
Katie Lim, Matthew Giordano, Theano Stavrinos +4
Direct-attached accelerators, where application accelerators are directly connected to the datacenter network via a hardware network stack, offer substantial benefits in terms of r…
cs.OS2024
Accelerator-as-a-Service in Public Clouds: An Intra-Host Traffic Management View for Performance Isolation in the Wild
Jiechen Zhao, Ran Shu, Katie Lim +4
I/O devices in public clouds have integrated increasing numbers of hardware accelerators, e.g., AWS Nitro, Azure FPGA and Nvidia BlueField. However, such specialized compute (1) is…