SAN JOSE, USA; San Jose, Multicore Expo: Netronome, a leading developer of network flow processors, has announced the Network Flow Engine (NFE-3240) family of PCIe 2.0 acceleration cards with enhanced I/O virtualization (IOV) support.
The NFE-3240 is the industry’s highest performance PCIe acceleration card, providing line-rate programmable L2-L7 packet and flow processing. The cards are designed for use by network and security appliances that require high performance and low latency, as well as highly-virtualized multicore servers.
The NFE-3240 improves the network performance of x86-based appliances and servers, providing a 10 times performance increase over standard network interface cards.
“Netronome’s NFE-3240 allows us to offer a world-class, real-time deep packet analysis solution to our customers, enabling them to measure unique speed and delay indicators in 3G/4G wireless networks,” said Bruce Peterson, CEO of Velocent Systems. “These capabilities enable us to itemize and prioritize traffic in order to ultimately solve the end-to-end QoS problems facing wireless operators.”
Delivering innovative acceleration for network and security apps
Powered by Netronome’s award-winning 40 Gbps NFP-3240 network flow processor, the NFE-3240 acceleration cards are available with two 10-gigabit Ethernet or six 1-gigabit Ethernet ports.
The NFE-3240 utilizes several techniques to dramatically improve network I/O workloads including packet classification, stateful flow analysis, deep packet inspection, integrated security processing and dynamic load balancing of flows across a virtualized PCIe datapath to multiple x86 CPU cores.
The NFE-3240 is ideal for line-rate packet capture or inline applications requiring zero packet loss and low latency. Linux-based applications can quickly benefit from a heterogeneous multicore architecture that couples the application processing power of multicore x86 with the high-throughput networking capabilities of the NFE-3240.
The heterogeneous multicore architecture allows the NFE-3240 to remove networking workloads that degrade x86 performance, enabling the x86 to provide more resources to application and control plane processing. Benefits of the NFE-3240 include:
* Programmable L2-L7 flow processing – offering 1,800 instructions per packet at 30 million packets per second, delivered by 40 multithreaded, RISC microengine cores running at 1.4 GHz.
* Line-rate networking performance – for packet capture, flow processing and deep packet inspection across all packet sizes.
* Integrated security processing – for 20 Gbps of line-rate cryptography and 20 thousand PKI operations per second.
* Enhanced IOV support – offering bandwidth guarantees and traffic isolation, with extremely low latency, to over two thousand x86 endpoints and other networking devices.
* Comprehensive development tools – including easy-to-integrate C-based APIs or a fully integrated development environment that includes a C-compiler, powerful simulation environment, extensive software libraries and sample applications.
* Energy efficient design – offering energy savings from the industry’s highest BIPs (billion instructions per second) per watt, providing over 70 percent reduction in power consumption over existing designs.
“Internet traffic continues to rise at dramatic rates and manufacturers of network and security appliances aim to keep pace, while offering greater levels of intelligence,” said Niel Viljoen, founder and CEO of Netronome.
“Combining the NFE-3240 with multicore x86 processors, designers can scale their applications to 40 Gbps and beyond while maintaining the benefits of standard x86 systems.”
Netronome’s NFE-3240 acceleration cards are available now with a range of price and performance options.