About a-team Marketing Services
The knowledge platform for the financial technology industry

A-Team Insight Blogs

NovaSparks Cites Nanosecond Performance of FPGA Data Appliance

Subscribe to our newsletter

NovaSparks has released performance data for its FPGA-based data feed appliance based on the highly volatile market conditions in August, noting average latencies of 400 nanoseconds, and extreme cases of 900 nanoseconds.

On test was NovaSparks’ Gen2 appliance, which is based entirely on FPGA (Field Programmable Gate Array) technology (i.e. unlike some competing products, no traditional CPU is used).

Gen2 is a 2U rackable appliance based on FPGAs from Altera, equipped with 21 ports for inputs and outputs, which can be configured for either 10 gigabit Ethernet or PCI-Express (up to 12 ports).  The appliance draws 180 watts of power.

10 gigabit data feeds from Nasdaq, NYSE Arca and Bats are currently supported, and a single appliance can handle multiple markets being fanned-out to several receiving servers.

NovaSparks’ tests focused on handling the Nasdaq TotalView full order book feed, and resulted in average processing latencies within the appliance of 400 nanoseconds, 99th percentile latencies of 700 nanoseconds, with a few extreme cases (e.g. long messages at peak burst times) of 900 nanoseconds.

In a real-life configuration, end-to-end latencies (from data source to receiving server) are typically a maximum of 1.5 microseconds, using PCI-Express as the interface. 10gE latencies are a bit slower, depending on the type of network interface card.

NovaSparks CEO Yves Charles says the company has both customers and proof-of-concept prospects using Gen2.  Customers and prospects are generally using the appliance for:

– Provision of managed services (such as Burstream)
– High Frequency Trading operations
– Market making activities
– Enterprise data delivery with a general low-latency requirement

Next up for Novasparks is to add further data feeds, with a focus on the options and futures markets.

Subscribe to our newsletter

Related content

WEBINAR

Recorded Webinar: Navigating the Build vs Buy Dilemma: Cloud Strategies for Accelerating Quantitative Research

For many quantitative trading firms and asset managers, building a self-provisioned historical market data environment remains one of the most time-consuming and resource-intensive steps in establishing a new research capability. Sourcing data, normalising symbologies, handling corporate actions and maintaining infrastructure can take months and absorb significant budget before a single model is tested. At the...

BLOG

ICE Adds Institutional Bond Execution with $6 Billion MarketAxess Acquisition

Intercontinental Exchange has agreed to acquire MarketAxess Holdings for $167 per share in cash, valuing the electronic bond trading platform at approximately $6 billion in equity and $5.7 billion in enterprise value. The price is a 33% premium to MarketAxess’s closing price on 29 July. MarketAxess connects around 2,100 institutional investors and broker-dealers in more...

EVENT

TradingTech Summit New York

Our TradingTech Summit in New York is aimed at senior-level decision makers in trading technology, electronic execution, trading architecture and offers a day packed with insight from practitioners and from innovative suppliers happy to share their experiences in dealing with the enterprise challenges facing our marketplace.

GUIDE

Connecting to Today’s Fast Markets

At the same time, the growth of high frequency and event-driven trading techniques is spurring demand for direct feed services sourced from exchanges and other trading venues, including alternative trading systems and multilateral trading facilities. Handling these high-speed data feeds its presenting market data managers and their infrastructure teams with a challenge: how to manage...