This section of the user guide describes how to use the ExaNIC as a developer. There are at least three different APIs that can be used depending on requirements:

  • Existing sockets applications can be accelerated with the exasock TCP/UDP acceleration library. Generally this is simply a matter of running the application with the exasock command. For more details and known limitations, see the exasock usage guide.

  • Sending and receiving packets with low latency using the ExaNIC API library (libexanic). Refer to the internals primer for an introduction to the architecture of ExaNIC cards, and the libexanic usage guide for API details. It is possible to use libexanic simultaneously with exasock (e.g. receiving UDP frames with libexanic and sending TCP frames with exasock), and it is also possible to use exasock to pre-build TCP frames and then send them using libexanic (see the TCP acceleration section).

  • Writing your own logic to run on the ExaNIC card's FPGA using the ExaNIC FDK. For more information refer to the FDK documentation. Note that the FDK is an optional component that is licensed separately. When running a custom firmware image, all network interfaces are still exposed to software (subject to filtering by the user FPGA logic), so the normal software APIs are still available.