Advantech Pumps Up 100G Solutions at Light Reading

Advantech to address the challenges of 100G traffic handling and its impact on SDN and NFV infrastructure. Peter Marek Senior Director of x86 Solutions at Advantech Networks and Communications Group, will discuss the need for a new platform design approach to efficiently handle 100G pipes.

Santa Clara, CA, Light Reading Next Generation Networking Components Conference, November 6, 2014 – Advantech (TWSE: 2395.TW) today are showcasing  two new products for 100G traffic handling and in a conference  panel session will discuss how to better address the SDN and NFV deployment needs of equipment providers and carriers.  Although white box server models bring economies of scale, a one-size-fits-all approach reaches its limits with multiple 40GbE and 100GbE ports caused by fundamental compute and I/O bandwidth constraints. Advantech will demonstrate how its new platforms offer the flexibility and performance needed for 100Gbps packet processing in next generation networks.

Advantech are stressing that on 100Gbps ports, virtual network functions do not scale as well on commodity hardware leading to cost inefficiencies through waste of data center or central office space and excessive energy consumption. As no single server can handle the millions of flows embedded in a single 100GbE pipe efficiently, distributing network traffic will become an issue in a virtual infrastructure.

“While some players may claim that the packet-level distribution schemes based on a few thousand of rules in a regular top of rack switch is sufficient, we strongly believe that flow distribution based on L4 to L7 criteria with millions of rules is what is needed to fulfil the promises that virtual infrastructure holds for service providers and carriers.

Panelists on the session “SDN & NFV: Assembling the Best Components” include participants from ARM, Advantech, Open Networking Foundation and Tieto PDS, and will be moderated by Simon Stanley, analyst at large for Heavy Reading.

The first 100G product to be showcased at the event is Advantech’s FWA-6512C Ultra Ultra-High-End Network Appliance with FlowNIC technology for 100G flow processing. The platform is a highly integrated appliance designed to boost performance where I/O density and processing power need to be maximized especially for network security, deep packet inspection and traffic/policy management in non-virtualized and virtualized environments such as SDN and NFV. Designed for handling the increased data rates in next generation networking infrastructure it supports connectivity of up to four 100GE ports in a compact 2U footprint. While offering performance levels that previously required more complex and costly architectures, Advantech’s FlowNIC technology integrates seamlessly with standard software frameworks such as Linux, Open vSwitch (OVS) and Intel® Data Plane Development Kit (DPDK).

To meet next generation networking needs at multiple 100G speeds, Advantech defined the platform from the ground up. The FWA-6512 series is architected around four Intel® Xeon® Processor E5-4600 v2 CPUs with support for SKUs up to 12 cores ( 48 cores total) and 1TB of system DRAM. The system incorporates a FlowNIC based on Netronome’s NFP-6240 silicon that acts as an intelligent load balancer between the external networking ports and the integrated Intel® Xeon® processors. Unlike conventional network appliances that use the QPI interfaces of the processors to route traffic between cores and virtual cores on different sockets, the FlowNIC inside the FWA-6512C is connected to the four Intel® Xeon® processors using four distinct PCIe x8 Gen3 connections, which enables network traffic to be delivered to the destination cores with optimized throughput and minimum latency.

The second 100G product on display at the event is Advantech’s new 100G AdvancedTCA hub blade.  The design brings together a unique combination of Broadcom switches with onboard Intel® Atom™ C2000 processor and two Virtex-7 690T FPGAs to enable a highly flexible and programmable hub blade with specialized packet processing capabilities on 100GbE ports

The ATCA-9223 targets network equipment providers delivering ATCA-based solutions for applications such as high performance network security in carrier and large enterprise data center networks, next generation 4G/LTE network infrastructure or high-capacity and high port density DPI-based  policy enforcement devices.

The hub blade is also aimed at SDN and NFV deployment where its wirespeed packet classification and DPI capabilities can be utilized for optimizing traffic flows and the load balancing of data across clusters of ATCA node blades.  Equipment providers can leverage Advantech’s recently introduced ATCA node blades based on the Intel® Xeon® Processor E5-2600 v3 series specifically optimized for Virtualized Network Functions.

Advantech will be showcasing the new products at the Light Reading Next Generation Networking Components Conference in Santa Clara at the Hyatt Regency on November 6, 2014.

For more information please contact nc*@ad*******.com or visit www.advantech.com/nc.