San Jose-based adaptable chip manufacturer Xilinx announced an update to its line-up of SmartNICs today, offering a SmartNIC platform that delivers convergence of network, storage and compute acceleration functions on a single device.
SmartNICs could be considered as intelligent Network Interface Cards that are targeted at the ever-changing and scaling needs of cloud data centers and high-performance compute applications.
The Alveo U25 SmartNIC is designed to bring the greater efficiency and lower TCO benefits of SmartNICs to cloud service providers, telcos, and private cloud data center operators. The U25 combines an optimized SmartNIC platform with a powerful and flexible FPGA-based engine that supports full programmability and turnkey accelerated applications.
In addition, Xilinx announced its first XtremeScale Ethernet adapter card in the Open Compute Project (OCP) Spec 3.0 form factor and a proof of concept for the first FPGA-based OCP Accelerator Module (OAM).
The Alveo U25 SmartNIC platform provides plug-and-play capabilities that make SmartNICs accessible for more widespread deployments.
Powered by Xilinx’s FPGA technology, the Alveo U25 SmartNIC provides higher throughput and a far more adaptable engine than SoC-based NICs. The platform enables ‘bump-in-the-wire’ network, storage, and compute offload and acceleration functions for efficiency by avoiding unnecessary data movements and CPU processing. This reduces the CPU burden and reclaims resources to run more applications. Embedded ARM processors provide control plane processing to support emerging bare metal server use cases. The baseline NIC delivers ultra-high throughput, small packet performance, and low-latency. Standard full-featured NIC functionality and drivers, including Onload application acceleration software, can reduce latency up to 80 percent and improve transmission control protocol (TCP)-based server application efficiency by up to 400 percent in cloud-based applications.
The U25 SmartNIC supports turnkey applications from both Xilinx and independent software vendors. The programming model supports high-level network programming abstractions such as HLS and P4, as well as compute acceleration frameworks such as the Vitis unified software platform to enable Xilinx and third party accelerated applications.
The first out-of-the-box accelerated application available on the Alveo U25 SmartNIC is support for Open vSwitch (OVS) offload and acceleration. The plug-and-play solution will offload over 90 percent of OVS processing from the server to improve packet throughput by over 5X. Future turnkey solutions from Xilinx are planned for security functions such as IPSec, SSL/TLS, AES-256/128, and distributed firewall as well as AI inference acceleration.
General availability of the Alveo U25 SmartNIC is expected in the third calendar quarter of 2020.
First OCP 3.0 Ethernet Adapter and OCP Accelerator Module
Xilinx also unveiled the new XtremeScale X2562 10/25Gb Ethernet adapter card based on the OCP Spec 3.0 form factor. Designed for high-performance electronic trading environments and enterprise data centers, the X2562 features sub-microsecond latency and high throughput with ultra-scale connectivity for real-time packet and flow information to thousands of virtual NICs. The X2562 is currently sampling and will be generally available in the second calendar quarter of 2020.
Additionally, Xilinx announed a proof of concept for the first FPGA-based Open Compute Accelerator Module (OAM). Based on the Xilinx UltraScale+ VU37P FPGA with 8GB of HBM memory and compliant with Open Accelerator Infrastructure (OAI), the mezzanine-based card supports seven 25Gbps x8 links to enable rich inter-module system topologies for distributed acceleration.