RDMA and the Linux IP stack

*

One Line Summary

Lets try to avoid using specialized tools and APIs for the RDMA stack and try to support the Linux network tools and APIs as much as possible.

Abstract

RDMA was developed as a kernel subsystem to drive highly specialized interconnects used in High Performance computing. As cloud computing developers to require more and more features for high speed interconnects the use of the RDMA subsystem becomes more popular. Sadly two network stacks have developed side by side often replicating in a different way how network interfaces are controlled and handled. This talk is a comparison between the tools of the standard network stack and the way that RDMA works and attempts to show areas where we could get closer to mainstream in order to ease the transition to RDMA from regular network I/O and to reuse portions of the standard stack as much as possible

Tags

posix network stack RDMA sockets APIs HPC management administration

Speaker

  • Biography

    Christoph Lameter is working as a lead in research and development for an algorithmic trading company in Chicago. and maintains the slab allocators and the per cpu subsystems. Over time he contributed to a number of Linux projects. As a kernel developer at SGI he helped pioneer the use of Linux for Supercomputing and developed the necessary kernel capabilities for HPC applications.