Talk @ Netdev 0x13: Improved System Call Batching for Network I/O

My colleagues Rahul Jadhav, Zhen Cao and me had our talk for Netdev 0x13 accepted. Our basic idea is reducing the CPU cycles consumed by amortizing the system call overhead of network I/O operations.

At the time we started our work, io_uring was not yet a thing. With the benefit of hindsight however, it is clear that io_uring is the way forward.

Here’s the video of Rahul delivering the talk:

And the slides: