Netronome unveiled its Network Flow Processing Platform,
which combines its latest generation of network flow processors with Intel Xeon 5600 CPUs to deliver 200 Gbps of programmable L2-L7 networking performance in compact 1U and 2U form factors.
Netronome's heterogeneous multicore processing architecture provides independent layers of workload-optimized packet, flow, security and application processing to dramatically increase network performance. Simultaneously, it frees IA/x86 resources to provide more application and control plane processing.
Because the platform is based on standard Intel-based CPUs, developers of Linux-based network and security applications can significantly reduce development time and costs. Highlights of the Netronome Network Flow Processing Platforms include:
- Application performance -- up to 200 Gbps of network and security application performance with sub-100 microsecond latency.
- Configurable network interfaces -- offering 200 Gbps of modular front-facing network interfaces with integrated network resiliency hardware including 10/100/1000, 10-Gigabit and 40-Gigabit Ethernet.
- System scalability -- through a 40 Gbps clustering module, traffic can be balanced across any number of devices for extremely high-throughput and compute-intensive workloads.
- Rapid development -- standard Linux-based applications can quickly and seamlessly take advantage of these capabilities through C-based APIs and a fully integrated development environment that includes a C-compiler, powerful simulation environment, extensive software libraries and sample applications.
- Programmable L2-L7 flow processing -- up to 480 Gbps of L2-L4 packet processing; line-rate, stateful L4-L7 flow processing; and accelerated security processing for bulk cryptography and PKI operations.
- Enhanced I/O virtualization -- for application and control plane traffic requiring bandwidth guarantees and traffic isolation in virtual environments, to over two thousand x86 endpoints and other networking devices.
