Software-Defined Networking (SDN) promises the vision of more flexible and manageable networks, but requires certain level of programmability in the data plane. Current industry insight holds that programmable network processors are of lower performance than their hard-coded counterparts, such as Ethernet chips. This represents a roadblock to SDN adoption.

In this paper we argue that contrast to the common view, the overhead of programmability is relatively low. We also argue that the apparent difference between programmable and hard-coded chips today is not primarily due to programmability itself, but because the internal balance of programmable network processors is tuned to more complex use cases. These arguments are backed with calculations and real-life measurements.

Read the conference paper