liburing VS onload

I've tinkered around with io_uring on and off for the last couple years. But I think it's really becoming quite cool (not that it wasn't cool before... :)). This was a really interesting post on what's new https://github.com/axboe/liburing/wiki/io_uring-and-networki.... The combination of ring-mapped buffers and multi-shot operations has some really interesting applications for high-performance networking. Hoping over the next year or two we can start to see really bleeding edge networking perf without having to resort to using DPDK :)

Thought I was doing something wrong at first, but after looking at examples and code, I just wasn't able to reach the epoll numbers. Looking on the Github page, there a few issues there with people who found the same thing, with their own examples. #1, #2

To address my own silly questions, yes, one should use the new fixed buffers described in this document: https://github.com/axboe/liburing/wiki/io_uring-and-networki...

We're working on this! https://github.com/axboe/liburing/issues/830

Link: https://github.com/axboe/liburing/wiki/io_uring-and-networking-in-2023

Other than that, I heard there's some 3rd party network libraries that are generally considered lower latency, one of them I believe either redhat uses for one of their systems, or it is sponsored by them: https://github.com/Xilinx-CNS/onload

I mean really you want to be doing some kind of (kernel network bypass with some sort of system programming language like C++ for low latency trading systems like this. Arb trading is usually not worth it unless you have a ton of expertise or some crazy mathematical edge that can offset. Plus you need to be guaranteed to execute the trade you ask for which most average joes don't have.

How about the x2 series from Xilinx?

It comes from solarflare who have a long pedigree of low-latency smartnics. They used to supply Cloudflare, and also supply like 50% of fintechs/financial markets.

You can also just use openonload to accelerate your programs. In this case just doing straight linux socket programming, which can be accelerated without dpdk. Or just use the generic linux driver if necessary.

https://github.com/Xilinx-CNS/onload

Ahki SingHania - Akhi is the engineering manager for the Execution team. He has a background in operating systems and networking. Before working on the Internet Computer, he worked on OpenOnload and Barrelfish.

https://github.com/Xilinx-CNS/onload any good?

It used to be openonload.org which seems to redirect to https://github.com/Xilinx-CNS/onload for the source to onload.