• Technical Conference:  30 March – 03 April 2025
  • Exhibition: 01 – 03 April 2025
  • Moscone Center, San Francisco, California, USA

SC452 - FPGA Prototyping for Optical Subsystems

Monday, 25 March
08:30 - 12:30 (Pacific Time (US & Canada), UTC - 08:00)

Short Course Level: Advanced Beginner

Instructor:

Noriaki Kaneda, Nokia, USA
Robert Elschner, Fraunhofer HHI, Germany

 

Short Course Description:

The course is intended for the students and engineers who have a basic background and experience in optical subsystems and the design of DSP algorithms but are beginners in the FPGA programming and FPGA prototyping of optical subsystems. The course gives insights to participants on FPGA prototyping by going through lecture materials and by a demonstration. The course will cover the following topics:

  1. Where are application areas for FPGA prototyping in optical subsystem design?
    1. “Classical” digital signal processing applications
    2. Machine-learning applications
  2. What can FPGAs currently do (types of resource blocks)?
    1. DSP cores, logic cores
    2. SoC (system on chip) processors
    3. SERDES (high-speed serial interfaces)
    4. Next gen. FPGAs
  3. What is a typical design flow for the implementation of DSP in FPGAs?
  4. How to adapt DSP processing blocks for FPGAs
    1. Floating-point vs. fixed-point processing
    2. Serial vs. parallel processing
    3. Pipelining
    4. Implementation in FPGA – best use of resources
    5. Examples from single-channel coherent transceivers & machine learning
  5. How to realize FPGA prototypes for optical subsystems
    1. Setup of required FPGA hardware
    2. Setup of hardware interfaces (DAC, ADC, SERDES)
    3. Demonstration of an FPGA-based coherent transceiver prototype


Demonstration:

The demonstration will take about 1 hour of the course time and comprises a full design cycle of FPGA programming: In the first part, a VHDL core is written from the scratch and simulated using open-source software. In the second part, the instructors will demonstrate the FPGA implementation of this VHDL core using a commercial design tool and test it on an on-site FPGA. Finally, the FPGA will be used in a demonstration of a full FPGA-based coherent transceiver prototype.

Short Course Benefits:

The course should enable participants to

  • Identify key applications and approaches of FPGA prototyping in optical subsystems.
  • Describe the key functionalities and capabilities of FPGAs for intended prototyping applications.
  • Define the difference between offline DSP and real-time DSP for FPGAs
  • Design the architecture to realize selective DSP functionalities.
  • Define the workflow of FPGA projects for implementation ready bit files.
  • Explain the design flow of implementing embedded SoC in FPGA for optical subsystems prototyping.
  • Describe the software and hardware architecture required to interface FPGAs and data converters (ADCs and DACs) for subsystem prototyping.
Short Course Audience:

The course is intended for students and engineers who have some basic background and experience in optical subsystems and the design of DSP algorithms but are beginners in the FPGA programming and FPGA prototyping of optical subsystems. The course gives insights to participants on FPGA prototyping by going through lecture materials and by a hands-on experience. Most of the materials are related to FPGA prototyping of Digital Signal Processing (DSP) in coherent optical transceivers, but will also cover aspects of prototyping for machine learning applications.

Instructor Biography:

Noriaki Kaneda works in CTO and strategy office of network infrastructure business of Nokia in Murray Hill, NJ. He was with Nokia Bell labs for over 20 years before he worked for nEYE systems in Berkeley, CA for 2021-2023. His expertise includes high speed DSP algorithms and implementations including machine learning based techniques for passive optical networks, short reach optical interconnects and long-haul coherent detection. He has done numerous FPGA projects for optical subsystem prototyping both in optical communications and Lidar systems, and many have been reported in his more than 90 publications in journal and conference papers. He also has received over 35 U.S. and international patents in the field of optical networking, high-speed DSP and microwave electronics.

 

Robert Elschner is a project manager and senior researcher in the department “Photonic Network and System” at the Fraunhofer Heinrich Hertz institute in Berlin, Germany. He joined Fraunhofer HHI in 2010 after receiving his Ph.D. from Technical University of Berlin. His research topics include digital coherent transmission technology for high-speed fiber-optical and THz-wireless systems, in particular the design of DSP algorithms and their real-time implementation on FPGAs. He has more than 100 peer-reviewed publications on these topics and was recently serving as a Subcommittee Chair on the Technical Program Committee of the Optical Fiber Communications Conference.