256 high-density QSFP-DD connectors yield exponential low latency edge ports, FPGAs
Arista Networks (NYSE: ANET) has this month introduced a line of converged ultra low latency, highly programmable 7130 Series systems, optimized for demanding in-network applications. The manufacturer says the new systems improve customer agility, while consolidating multiple devices that reduce complexity, power and costs.
The 7130LBR Series
The 7130LBR Series are 96 port systems combining Arista's ultra low latency Layer 1 X-Point platform with two high performance FPGAs, Broadcom Jericho2 switching silicon, a powerful CPU and precision timing hardware, enabling consolidation of network, server and custom FPGA applications into a highly compact form factor.
Compared to traditional multi-box approaches, the manufacturer contends the 7130LBR reduces device and network hop count, cabling, power and cost, while providing orders of magnitude density and latency improvements through co-location of these key components.
Supporting dynamic reconfigurability, the robust features of Arista’s R series, a portfolio of 7130 FPGA applications from Arista and its partners, and a comprehensive FPGA development kit, with the new system, the company says customers can mix and match Arista applications with their bespoke processing without resorting to slow connections to external devices.
In addition, two high-grade AMD/Xilinx FPGAs can be combined to perform complex data processing tasks, or deployed in parallel enabling independent functions and live prestaging of new applications with seamless switch-over.
The 7130B Series
The 7130B Series provides high density Layer 1 fabric capability combined with a full-featured, high-radix Intel Tofino L2/3 switch.
Providing over 2.5x the Layer 1 connectivity of existing solutions with 256 interfaces using high-density QSFP-DD connectors, with the new system, Arista maintains that hundreds of devices can be interconnected to provide thousands of edge ports and hundreds of FPGAs with minimal latency impact.
In addition, the 7130B model's sophisticated P4 programmable packet processing enables customized packet handling, filtering and statistics, as well as Arista’s advanced pipeline profiles for scalable address translation, tunneling and load balancing.
According to the manufacturer, together, the 7130LBR and 7130B enable a new dimension of composability, performance and scale for high performance FPGA-native applications for financial services, network monitoring, government and broadcast applications.
Key options, features and testimonial
According to Arista, via the 7130 Series, integrating full-featured, programmable switching silicon from industry partners with ultra low latency Layer 1 and highly scalable FPGAs on a single platform allows financial market participants to compose multiple trading logic elements in highly scalable, low latency switching fabrics.
The expanded 7130 Series of low latency programmable switches includes two new models that integrate full-featured L2/3 switching and open programmability with high performance L1 connectivity. They both run Arista EOS and integrate with the company's CloudVision platform, supporting a broad range of switching, routing, DANZ monitoring and programmabilit functions across Arista’s extensive portfolio.
Key features of the 7130 Series, per Arista include:
- Reduces network latency to support the increasing demands of financial trading solutions.
- Dynamically composes key FPGA, networking and software applications within a single tier to increase application performance and reduce complexity.
- Consolidates servers, network and FPGA devices to reduce footprint, power and cost.
- Scales out applications to support larger numbers of clients, venues or data streams.
“Cboe is one of the world’s largest global market infrastructure providers and delivers leading-edge technology that powers markets and serves investors around the globe,” asserted Curt Schumacher, SVP, Infrastructure at Cboe Global Markets. “As we continue to invest in our infrastructure to drive our ongoing expansion into new markets and regions globally, we expect Arista’s solutions will provide new ways to further enhance application performance and scale, without compromising on efficiency, in our business.”
Application partnerships
The next generation 7130 platforms offer compatibility with Arista’s portfolio of FPGA applications, including solutions for finance, such as MetaMux and ExchangeApp, as well as high-capacity traffic capture and pico-second accurate timestamping with MetaWatch. Further, they system's integrated high performance switching silicon leverages the rich capability of EOS, including the switching and routing suite, packet broker capabilities of DANZ for network monitoring and the P4 applications for the Tofino packet processor. Finally for developers, Arista’s EOS provides a mature programming environment with API, FDK, and SDK and full support for AMD/Xilinx’s Vitis Unified Software Platform, Enyx’s nxFramework and Intel’s P4 Studio for Tofino.
According to Arista, this comprehensive combination of processing capabilities ensures the next generation 7130 family is an ideal fit for applications that currently require separate dedicated appliances or server-hosted processing modules in finance, government, broadcast media and other mission-critical deployments.
Availability
The new additions to the 7130 Series are fully integrated with Arista EOS and CloudVision for consistency in day-to-day operations. They are packaged with a comprehensive FDK to support custom application development. Customer testing is currently in progress, with general availability in Q4 2022.