DPDK

Data Plane Development Kit (DPDK) is a bundle of libraries and controllers that helps us to build applications, and thanks to which commodity hardware can achieve higher throughput to software. In short, it is a tool that unlocks faster packet processing. This is achieved by bypassing Kernel space entirely, running everything in user space. It was designed by Intel to be integrated with the x86 line of processors. Various industries can benefit from DPDK deployment, since it has already become the de facto standard in the NFV and SDN world. It can accelerate virtual applications like next-gen firewalls, applications for Big Data analysis, or switching, while Open vSwitch is a prime example of a successful DPDK application. 







dpdk-logo.png

Even though DPDK is a powerful tool for data plane processing, not every network interface card supports it. Fortunately, Netcope FPGA-based network adapters come with native support for DPDK and push the limits of DPDK even further by combining Netcope Data Plane (NDP) with DPDK. Since the maximal theoretical throughput of PCI Express is much lower than the maximal theoretical memory throughput of the x86 line of current server processors, NDP has been designed to maximize throughput between FPGA-based network interface cards and the host system. In other words, NDP squeezes as much bandwidth as possible from the PCI Express interface, mitigating the bottleneck of PCI Express limited throughput.

NDP forwards packets through the card and PCI Express in a clever and efficient manner. NDP and DPDK work together as a layered solution. The lower layer (NDP) forwards data through the PCI Express, and the upper layer provides access to data adhering to the DPDK API and data model. With this solution it is possible to achieve full duplex 100 Gbps throughput for packet lengths above a certain threshold and near 100 Gbps throughput for short packets. Simply put, DPDK accelerated by NDP is a perfect fit for you if you need full DPDK API compatibility as well as very high throughput, and you plan to use a high-performance CPU for your task.

Download our whitepaper and learn how we are able to achieve 148 Mpps throughpout, a maximum theoretical throughput for DPDK.
 

Whitepaper 
 

Explore more Netcope products

Netcope Packet Capture (NPC)

Netcope Packet Capture (NPC)

Learn more about the network solution for wire-speed, low-latency packet processing.

Netcope Session Filter (NSF)

Netcope Session Filter (NSF)

Learn more about Netcope's powerful session-oriented packet capture solution.

Netcope FPGA Boards (NFB)

Netcope FPGA Boards (NFB)

High-performance appliances designed for the development of hardware-accelerated applications.

Netcope Success Stories

NSF-100G2-Picomass-Netcope

NSF-100G2-Picomass-Netcope Network Traffic Monitoring

Picomass uses Netcope Session Filter in the IPS200 DPI solution because NSF is capable of offloading traffic to hardware. This makes real-time DPI on 100G networks possible.

Flowmon Networks

Flowmon Networks Network Traffic Monitoring

Flowmon Networks aims to develop a new generation of NetFlow/IPFIX probes that can monitor and process traffic of 100GE high-speed networks. Because of that, Flowmon Networks needs to find a network card that would be able of handling 100 GE monitoring requirements.

U.S. trading firm deploys Tradecope

U.S. trading firm deploys Tradecope Electronic Stock Trading

For successful trading on electronic exchanges today, it is not enough to come up with the smartest strategy anymore. Learn more about how U.S trading firm deployed FPGA-based Tradecope solution to increase hit rate of the trading strategy.

Cookies help us deliver our services. By using this website, you agree to the use of cookies.  More information

close