partner-offering-banner.png

NVMe Over Fabric (NVMe/TCP) 25GbE Initiator IP (NVMeTCP25G-IP)

DesignGateway Co., Ltd.

Select

As cloud computing continues to evolve rapidly, the need for high-speed, low-latency storage access has become essential for enterprises handling massive volumes of data. Traditional CPU-based cloud storage solutions often face challenges such as bottlenecks, reduced power efficiency, and limited scalability, which can hinder overall system performance. By offloading NVMe/TCP processing to an FPGA-based engine, these limitations can be overcome. The FPGA architecture significantly reduces CPU load, minimizes latency, and enhances data transfer efficiency, resulting in a dramatic improvement in cloud storage performance. While many cloud infrastructures rely on standard Network Interface Cards (NICs) to handle NVMe/TCP connections, CPU-based processing introduces inherent inefficiencies. An FPGA-based NVMe/TCP offload engine eliminates these bottlenecks, delivering a more scalable, power-efficient, and high-performance storage solution for modern data centers.

Key Features

  • CPU-less NVMe/TCP initiator IP core enabling remote NVMe storage access directly from FPGA
  • 128-bit Avalon-MM bus for data interface
  • Configurable 4KB command depth: Up to 128
  • High-performance 4KB data transfer: Up to 2500 MB/s write access, 2600 MB/s read access
  • Easy integration for rapid development and focus on core application logic
  • Scalable storage architecture—performance and capacity can grow with the server, not the FPGA
  • Complete hardware platform: DYNANIC DMA driver & hardware engine, supporting virtualized environments
Expand Close
Keyfeatures main image

Offering Brief

Offering Brief

Device Family Agilex® 7 FPGAs and SoC FPGAs F-Series
Offering Status Production
Integrated Testbench No
Evaluation License No
Design Examples Available Yes
Demo Yes
Compliance No
Latest Quartus Version Supported 23.1.0
Development Language Encrypted VHDL

Encrypted IP core

Reference Designs Quartus Project

Ordering Information

Market Segment and Sub-Segments