NVIDIA Accelerated IO (XLIO) Documentation Rev 3.61

Overview

The NVIDIA® Accelerated IO (XLIO) software library accelerates TCP and UDP network applications by delivering high bandwidth, low latency, and reduced CPU utilization. Built as a user-space library with a kernel-bypass architecture, XLIO enables direct data transfer between application memory and network adapters for optimal performance.

XLIO provides two APIs to meet different application requirements:

  • POSIX Socket API – offers zero-cost integration with existing applications. No code changes are required to achieve substantial networking performance improvements; applications can be accelerated simply by preloading the XLIO library.
  • Ultra Socket API – designed for applications demanding maximum performance. While it requires additional integration effort, it enables advanced capabilities such as true zero-copy data transfers and a simplified, highly optimized data path.

XLIO seamlessly integrates with cryptography-enabled NVIDIA® ConnectX® network adapters and NVIDIA® BlueField® DPUs, providing hardware-offloaded Transport Layer Security (TLS) symmetric encryption and decryption. It also leverages hardware acceleration features such as Large Receive Offload (LRO), TCP Segmentation Offload (TSO), and Striding-RQ to further enhance performance for both POSIX and Ultra API users.
XLIO is an open-source project, and we welcome contributions from the community. Please visit us on GitHub and see our Contributing Guide to get started.

Document contents:

System Requirements

The following table presents the currently certified combinations of stacks and platforms and supported CPU architectures for the current XLIO version.

SpecificationValue
Network Adapter CardsNVIDIA Bluefield-3, 200 GbE, Dual Port, 32GB RAM
NVIDIA ConnectX-7, 200 GbE, Dual Port
NVIDIA ConnectX-6 Dx, 100GbE, Dual Port
FirmwareBlueField-3 32.47.1026
ConnectX-7 28.47.1026
ConnectX-6 Dx 22.47.1026
Driver Stack
  • DOCA-Host v3.3.0

    • doca-libxlio profile
    • doca-roce profile

NOTE : Starting this version, the host driver is part of the NVIDIA DOCA package.

For further information, please see NVIDIA MLNX_OFED to DOCA-OFED Transition Guide .

Supported Operating Systems and Kernels
  • Ubuntu 22.04
  • X86 only:

    • Ubuntu 24.04
    • RHEL 9.6
  • DPU only:

    • CTYUNOS 23.01
CPU Architecturex86_64, Arm (on NVIDIA BlueField-3 and NVIDIA GH200 Grace Hoppera)
Minimum memory requirements1 GB of free memory for installation 800 MB per process running with XLIO
Minimum disk space requirements1 GB
TransportEthernet

a. NVIDIA GH200 Grace Hopper is only supported with Ubuntu 22.04.

Document Revision History

For the list of changes made to this document, refer to User Manual Revision History.

© Copyright 2026, NVIDIA. Last updated on Feb 3, 2026