SANTA CLARA, Calif.
April 2, 2013
/PRNewswire/ -- Algo-Logic Systems, Inc., a recognized leader in providing hardware-accelerated, deterministic, real-time, ultra-low-latency products, systems and solutions for accelerated finance, packet processing and embedded system industries, announces that it will present live demonstration of its newest FPGA-accelerated Market Data Filtering (MDF) application. This live demo will be showcased on
Monday, April 8th, 2013
at the HPC Wall Street Conference at Roosevelt Hotel in
New York City
Algo-Logic's MDF application deterministically filters market data to eliminate CPU queuing delays that traders encounter during micro-bursts and peak market activity. The MDF application receives UDP/IP market data from multicast groups and outputs UDP/IP packets containing only the relevant fields of interest parsed from the filtered messages. The MDF comes with a Graphical User Interface (GUI) to easily control key hardware parameters and chart data filtered by the FPGA. The Application Programming Interface (API) can be customized to augment existing Linux-based software trading solutions with FPGA hardware. The MDF operates at full 10 Gbps line rate and incurs less than 20 nanoseconds (ns) of market data parsing and filtering latency, less than 450 ns of total packet processing latency, and less than 1 µsec of pin-to-pin latency.
Booth Location #105:
Algo-Logic Systems will demonstrate the live Market Data Filtering hardware at booth #105. The live MDF demo runs on a Solarflare AOE card in a host running Linux Centos v6.4 receiving a NASDAQ ITCH 4.1 feed with 12 multicast groups. While the hardware parses and filters the messages, the software dashboard displays price history of a dozen sample symbols in color-coded charts highlighting discovered market conditions and trading opportunities.
The MDF application was implemented using Algo-Logic's pre-built low latency library. Algo-Logic's FPGA-based, hardware-accelerated, deterministic, and ultra-low-latency library and generic building blocks are ideal for lowering latency in co-located servers. Algo-Logic's low latency library includes components for feed parsing, risk checks, customized trading strategies, and order injection.