Anything from simple automation to an ultra-low latency trading strategy or high-performance 100G Ethernet switch can be implemented in FPGA. Depending on where the FPGA is placed you can have your system running in anything from FPGA enabled network card to industrial grade high-speed device. Moreover: you might want to create a card of your own, or a new industrial grade device, or both, and put it one into the other. To do any or all of it you will need FPGA experts. You will want us: LDA Technologies.
LDA Technologies is a team of leading professionals with over 20 years of experience in product design. Starting 2010 we have been concentrating on FPGA development and delivering FPGA-based products to our customers.
• FPGA-based systems design and programming
• Software: ultra-low latency, management, drivers and embedded
• Hardware Design: FPGA, PCB, Mechanical
• Testing and benchmarking
• Low-volume manufacturing
• Documentation and guides
• Regulatory certification
• Maintenance and Support
• Ultra-low latency, High Frequency Trading
• Networking and switching
• Packet processing
• Network security
• Test and measurement
• Digital and IP Video
• Small footprint microcontroller cores
• DDRx memory controllers
• Digital signal processing
• Power/signal integrity analysis
• Digital design
• Data Acquisition
One of the main services provided by LDA Technologies is Design and Development of FPGA-based applications: IP Cores and accompanying software when needed.
Choosing the right FPGA for the task at hand; choosing the right FPGA board for the IP core to run; designing the system as a whole: FPGA – firmware – software; designing/implementing algorithms; writing efficient FPGA and software code; testing and deploying: these are all parts of FPGA IP core creation process.
At LDA Technologies we have extensive expertise producing IP cores for many FPGA types and vendors: from ultra-cheap ones to high-end FPGAs capable of running most sophisticated trading strategies, packet processing and computational tasks.
Virtex 7, 6, etc.
All FPGA series starting 1995
V, IV, etc.MAX series
3rd party hardware we’ve been working with includes:
- Alpha Data
- HiTech Global
- Arista Networks
We have extensive experience of designing complex algorithms from scratch as well as being a building block in a bigger IP core: both incorporating 3rd party cores into ours and building an IP core in collaboration with other teams.
- Tamba Networks
- PLDA (PCIe cores)
- Northwest Logic
- Altera Quartus
- Xilinx Vivado
- Cadence Orcad, Allegro
- Mentor Graphics PADS PCB
- Mentor Graphics ModelSim, QuestaSim
- Microchip MPLab
- Synopsis tools
- Mentor Graphics Nucleus OS
- RedHat eCOS
- Windriver VXWorks
- Micrium uCos-III
Software tools and technologies
- Languages: C/C++, Java, Python, C#
- Parallel Computing: OpenCL, CUDA, Intel Xeon Phi
- Systems: Linux x64, Linux x32, Windows
- Tools: GCC 4.x and 5.x, GDB, Intel C/C++ Compilers, Microsoft Visual Studio, Eclipse
- Technologies: Multithreading, Ultra-low Latency Processing, Distributed Computing, High-Frequency Trading
Device Manufacturing (ODM)
In addition to designing and developing FPGA-based systems on 3rd party hardware, LDA Technologies has the capability to design and manufacture boards and devices ourselves. LDA Technology has years of experience in:
- Thermal analysis and simulation
- FPGA (if applicable)
- Software/hardware co-design
- Embedded design and development
- Application level development
We have capability, knowledge and contacts in place for manufacturing your device in North America (USA/Canada) or overseas (China, Taiwan) or both. Combined with our highly experienced and efficient engineering team, our partner network and testing infrastructure it produces the synergy effect that allows us to achieve unprecedented time-to market for the product delivery.
Find out more about turnkey products manufactured by LDA Technologies here .
LDA Technologies released many custom and exclusive products that are currently deployed in the major US exchanges and throughout the world.
High speed, ultra-low latency market feed handlers
LDA’s largest line of FPGA-based systems was created exclusively for a US-based company supplying IT services to private trading shops. In the period of 2010-2015 we created a huge set of custom ultra-low latency market feed handlers on Altera Stratix V and XILINX Virtex family FPGAs for:
- NASDAQ: ITCH4.x and ITCH 5.0
- NYSE: ARCA Book and ARCA XDP Integrated
- BATS: PITCH
- DIRECT EDGE: Next Gen Multicast Market Data (before merger with BATS)
- CME: MDP2 and MDP3
- Wire-to-user application latency significantly below 1 microsecond
- LDA created a unique multicore processing engine for market data processing, which allows to processand synchronize up to four messages simultaneously.
- All work is done on the card: feed parsing, order book construction, re-requesting of dropped packets
- Data bursts are handled at link speed thus providing deterministic performance
- Data processing speed capacity over 1Gb/s
- Comprehensive, ultra-low latency, small footprint C++ APIs:
- Normalized book is presented to the application
- Same API interface for the same asset classes: stocks and derivatives
- Allow handling FPGA output by multiple CPU cores
- Feed configuration differences handled via special configuration files
For one of its customers LDA Technologies designed and manufactured a whole product line of FPGA-based Layer 1/Layer 2 network switches. LDA designed the product from scratch and was responsible for complete lifecycle: design, manufacturing, FPGA and software code development, assembly and end-customer support. Three generations of the devices were introduced during two years and found their applications in multiple industries and data centers across the world:
- Gen1 device comprised a full mesh 48-port Layer 1 switch primarily used by trading firms to reduce the latency on downstream data from exchanges.
- Gen2 device built on the Gen1 by adding a 3rd party FPGA board to the backplane, to provide low-latency Layer 2 switching and filtering functionality on top of Layer 1.
- Gen3 device was a major redesign of the product, with reimagined hardware architecture, advanced storage subsystem and an advanced FPGA board designed in-house. Gen3 device, paired with innovative FPGA-based switching, packet processing and capture functionality, represented the best-in-class device in terms of performance and latency at the time it was produced.
Firmware and software features also evolved to a powerful set of FPGA-based switching and monitoring tools in Gen3, including:
- Ultra-low latency switching (muxing) of multiple 10G inputs into a single outputs
- Network traffic filtering based on set of rules
- High-performance network traffic capture, utilizing on-board SSD raid or an external storage
- Extensive health monitoring, using the industry-standard protocols (SNMP, Syslog)
- Set of proprietary customizations for some end-customers.