Layer 4 Load Balancer in eFX and Multi-asset Class Trading Case Study

Business Challenge

Large brokers require a networking infrastructure that has the reach and flexibility to underpin their client connectivity platform, to be able to access the order flow to drive their business.

A provider of client connectivity and distributed FX ECN platform for financial firms is looking to improve its network to keep up with the financial arms race, delivering on-demand, secure and scalable infrastructure.

While network switches and routers have evolved to support immense growth of data traffic, network service appliances and server capacity remain limited to a couple of hundreds of gigabits at best, far below the multi-terabit network throughput capacity. Considerable resources are invested in deploying of hardware and software load balancing for scale-out. Increased infrastructure costs and the burden of life cycle management for the load balancing infrastructure hold the business back.

Scale-out Load Balancers

Approach

Netberg Aurora 610, powered by the Tofino chip and P4 programmability, bridges the performance gap between multi-terabit switches and gigabit servers and appliances.

A custom fit-for-purpose software stack, based on Open Network Linux and Netberg-provided BSP for the P4 Studio SDE, took its place at the control plane.

Supporting multiple load-balancing mechanisms, resilient hashing, and flexible allocation of hardware resources to load-balance millions of connections, load-balancing can be done inside the switch, providing multi-terabit traffic distribution for Layer 3 and 4 services and applications.

Business Results

A large number of software-based load-balancer servers were replaced by Netberg Aurora 610 switches, reducing the cost of load balancing by order of magnitude, with a distributed architecture and optimized traffic path.
Tofino Load Balancers

Bringing modern Tofino-based Netberg switches in the path gave multiple benefits to the service provider:

  • The ASIC has resources and primitives to guarantee connection affinity even in the presence of concurrent DIP pool changes and millions of connections, while allowing for low latency, and terabit speed
  • Embedded endpoint health checks with zero runtime reliance on control plane
  • Guaranteed sub-second detection of pool member failure, and traffic redirection
  • P4 programmability enables developing of additional features, such as in-line security

 

Why Netberg

Netberg meets the growing demand for open network hardware based on merchant silicon.

  • A wide range of products with throughput rates from 1 to 100 Gbit/s.
  • Open Network Install Environment (ONIE)-enabled switches allow choice of networking software platform.
  • Support for Open Networking Linux, a part of the Open Compute Project and a component in a growing collection of open source and commercial projects including CORD and Stratum.
  • High-quality hardware.