Describes the design and implementation of the datapath classifier – aka dpcls – in Open vSwitch* (OVS) with Data Plane Development Kit (DPDK). It presents flow classification and the caching techniques, and also provides greater insight on functional details with different scenarios.
Configure the DPDK link aggregation feature. Link aggregation combines multiple network connections in parallel to increase throughput beyond what a single connection could sustain.
Part 2 of a series that describes design and implementation of the OvS-DPDK datapath classifier, dpcls. Call graphs and the subroutines involved are covered here.
This article shows how we used Intel® VTune™ Amplifier to identify and fix an MMIO transaction performance bottleneck at the microarchitecture level in OVS-DPDK.
Describes how to characterize NFVI latency and use optimization toolsets and fine tuning to achieve deterministic NFV performance in an open source infrastructure. Aligns with OPNFV best practice for tuning Linux-based KVM hypervisors.
The DPDK Packet Framework enables rapid prototyping of real-world workloads on multicore Intel® CPUs. Uses DPDK libraries librte_port, librte_table, and librte_pipeline.
Learn about use cases that include SDN/NFV packet processing, Hadoop* data movement, and accelerated data encryption.
Learn how to configure an SDN enabled wired/wireless network using Open vSwitch* (OvS) with Faucet*, which is an open source SDN controller. Detailed instructions are provided.
Memcpy is a simple operation implemented many ways. Here we describe the optimization approach used for memcpy in the Data Plane Development Kit (DPDK).