← Projects

Dataplane Batching

Implement dataplane batch sending

---

Details

Certain dataplanes allow for batch sending messages to them (linux kernel via netlink). With our dataplane abstraction in zebra/zebra_dplane.c, we create a queue of dataplane context objects for messages we want to send to the kernel. In a separate thread, we loop over this queue and send the context objects to the appropriate dataplane. A batching enhancement should tightly integrate with the dataplane context objects so they are able to be batch send to dataplanes that support it.

The dataplane netlink code in zebra/rt_netlink.c error handling will need to be adjust from what it is currently (they know an error occurred for message X immediately). With batching, you send N messages and maybe 1 of them fails. You will need to relate that failed message to a context object via a sequence of some kind and alert the upper level dataplane handling code about it.

Bonus Goal

The dataplane context objects are currently malloc()d/free()d every time they are created/handled. Switching these to use a memory pool scheme where the context objects are pre-allocated and handed off as needed would improve processing time and should be a logical extension of batching.

Quick Start

• Familiarize yourself with the dataplane handling in zebra/zebra_dplane.c
• Review how netlink kernel messaging works. Our code is in zebra/rt_netlink.c
• Take a look at libmnl’s implementation of netlink batching.

Mentors

Stephen Worley (@sworleys)
Quentin Young (@qlyoung)

Helpful Links

FRR dev guide
FRR user guide