Kmod-nft-offload [updated] -

The Power of Hardware Offloading: A Deep Dive into kmod-nft-offload

Data centers: kmod-nft-offload can help data centers optimize network performance, improve scalability, and reduce power consumption.
Cloud environments: Cloud providers can benefit from kmod-nft-offload by offering improved network performance and scalability to their customers.
High-performance computing: kmod-nft-offload can be used in high-performance computing environments, such as research institutions and financial institutions, where fast and efficient network processing is essential.

Further Reading:

Dependencies: To function correctly, it typically requires other kernel modules, including: kmod-nf-flow: Handles the core flow table logic. kmod-nft-offload

How does it work?

Implements the flow_offload infrastructure for nftables
Translates nft rules (matching on 5-tuple + actions) into hardware flow entries
Falls back to software processing for unsupported actions
Works with drivers that expose flow_block_cb API

By offloading nftables rules to hardware, kmod-nft-offload alleviates the CPU burden, allowing it to focus on more critical tasks. This results in improved network performance, characterized by increased throughput and reduced latency. The Power of Hardware Offloading: A Deep Dive

Users have reported significant speed differences depending on whether offloading is active. For example, on some hardware, enabling offload can increase speeds from ~200Mbps to over 600Mbps, though actual results depend heavily on your specific router hardware and firmware configuration. How to Use It Data centers : kmod-nft-offload can help data centers

Requirements

Linux kernel 5.4+ (recommended 5.15+ for stable offload)
Network driver with TC_FLOWER or nft_offload support (e.g., mlx5_core, bnxt_en, dpaa2, netdevsim for testing)
nftables ≥ 0.9.3

How to Use It (Configuration)

To use nft-offload, you define a flowtable in your nftables configuration and associate it with a specific network interface.