Leveraging High-Performance Data Transfer
Our ASCR SmartNIC project published the following unclassified unlimited release (UUR) paper (SAND2024-1022C) in IEEE Cluster.
Abstract
Network interface controllers (NICs) with general-purpose compute capabilities ('SmartNICs') present an opportunity for reducing host application overheads by offloading non-critical tasks to the NIC. In addition to moving computation, offloading requires that associated data is also transferred to the NIC. To meet this need, we introduce a high-performance, general-purpose data movement service that facilitates the offloading of tasks to SmartNICs: The SmartNIC Data Movement Service (SDMS). SDMS provides near-line-rate transfer bandwidths between the host and NIC. Moreover, SDMS's In-transit Data Placement (IDP) feature can reduce (or even eliminate) the cost of serializing data on the NIC by performing the necessary data formatting during the transfer. To illustrate these capabilities, we provide an in-depth case study using SDMS to offload data management operations related to Apache Arrow, a popular data format standard. For single-column tables, SDMS can achieve more than 87% of baseline throughput for data buffers that are 128 KiB or larger (and more than 95% of baseline throughput for buffers that are 1 MiB or larger) while also nearly eliminating the host and SmartNIC overhead associated with Arrow operations.
Publication
- IEEE Cluster Paper Scott Levy, Whit Schonbein, and Craig Ulmer, "Leveraging High-Performance Data Transfer to Offload Data Management Tasks to SmartNICs". in IEEE International Conference on Cluster Computing (CLUSTER), September 2024.