SPIE Digital Library Get updates from SPIE Newsroom
  • Newsroom Home
  • Astronomy
  • Biomedical Optics & Medical Imaging
  • Defense & Security
  • Electronic Imaging & Signal Processing
  • Illumination & Displays
  • Lasers & Sources
  • Micro/Nano Lithography
  • Nanotechnology
  • Optical Design & Engineering
  • Optoelectronics & Communications
  • Remote Sensing
  • Sensing & Measurement
  • Solar & Alternative Energy
  • Sign up for Newsroom E-Alerts
  • Information for:

SPIE Photonics West 2017 | Register Today

SPIE Defense + Commercial Sensing 2017 | Call for Papers

Get Down (loaded) - SPIE Journals OPEN ACCESS


Print PageEmail PageView PDF

Optoelectronics & Communications

Optical interconnects for satellite payloads: sizing up the state of the art

New optical interconnects have the potential to perform in harsh environments like outer space, where future digital satellite systems will require high data-communication speeds and low power consumption.
30 April 2010, SPIE Newsroom. DOI: 10.1117/2.1201003.002685

Broadband communication services such as high-definition TV, video on demand, and Triple Play are fueling the bandwidth appetite of individual subscribers and network-service providers. Satisfying this hunger requires a steady increase in total available bandwidth for terrestrial networks, most effectively using optical fiber. In fact, today optical fiber is the main trunk for all data traveling long distance in wide- and local-area networks. Furthermore, individual customers can benefit from the huge data capacity of optical fiber by deployment of fiber to the home. Operators in the satellite-telecommunication sector face similar challenges driven by the same bandwidth demand as their terrestrial counterparts. Moreover, the limited number of orbital positions for new satellites has spurred an increase in payload data-communication capacity using an ever-increasing number of complex, multibeam active antennas and greater aggregate bandwidth. Only satellites with very large capacity, high computational density, and flexible, transparent, fully digital payload solutions can achieve affordable communication.

To keep pace with these requirements, communication-satellite designers must devise new systems requiring a total digital throughput of a few terabytes per second (Tb/s), resulting in a high-power-consuming satellite payload. An estimated 90% of the total power consumption per chip is used for off-chip communication. These factors taken together cause signal distortions in high-speed electrical data communication. For this reason, the European Space Agency commissioned us and others to carry out a literature study of interchip optical communications and photonic printed-circuit boards (PCBs) for next-generation onboard processors. Optics has proved immune to electromagnetic interference, and at intrasatellite interconnect lengths—between 50cm and a few meters—the optical medium itself does not impose any bandwidth limit. The question is whether the complete optical channel, from laser driver to detection circuit, can cope with the bandwidth demand and power constraints typical at the inter- and intraboard levels.

Research into board-level optical interconnects accounts both for the majority of papers in the field of investigation and for a large increase in publications in recent years (see Figure 1). Closer examination of the evolution of the different board technologies reveals an emphasis on waveguiding (see Figure 2). So-called optical wireless solutions—an approach whereby a sender addresses multiple receivers in free space—represent a minority of the literature. Most papers dealing with rack-to-rack interconnects focus on either ribbonized systems or imaging fiber. Finally, on-chip optical interconnects have been dominated by silicon photonics from 2006 onwards.

Figure 1. Growth in number of papers on board-level interconnects.

Figure 2. Growth in number of papers on board-level optical-interconnect technologies.

To date, a variety of optical elements and guiding structures for coupling signals between optoelectronic transceivers have been proposed and developed. Figure 3 shows examples of intrachip and interboard interconnect applications, together with commonly studied optical-media technologies. The main research themes include free-space optical interconnects, fiber ribbons or embedded optical fibers, and embedded optical waveguides. In view of the special requirements for applying optical interconnects in space, only very robust, lightweight, hermetically sealed, and radiation-hardened solutions are suitable. We have determined that a guided-wave approach, using either fibers for near-term implementations or embedded waveguides over the longer term, have the best potential for deployment in space.

Figure 3. Schematic illustration of different kinds of optical media for board-level interconnects. MCM: Multichip module.

Undoubtedly, the two most important parameters are power consumption and aggregate bandwidth. The group of Cook1 published a comprehensive overview of typical fiber-based transceiver modules with respect to aggregate bandwidth and the number of available communication channels. We have updated their findings by adding the latest research on short-reach optical interconnects (see Figure 4). The total aggregate data rate has grown considerably. Record-setting aggregate bandwidths as well as total power needed to send one bit over a communication channel were reported by the IBM Watson Research Center (see Figure 4).2–4 In particular, the group of Shares4 found that very high aggregate bandwidths (480Gb/s) over board-level distances (30cm) do not necessary imply high power consumption (<5mW per Gb/s per channel) when implementing optical interconnects.

Figure 4. (top) Update on progress in board-level optical-interconnect research. (bottom) Power per bit versus reported aggregate data rate.

In conclusion, optical interconnects have a clear advantage over their electrical counterparts vis-à-vis power dissipation and aggregate bandwidth for board-level interconnects and beyond, i.e., interconnect lengths of 60cm and above.5–7 The main limitation is the opto-electronic conversion data rate. For very short interconnect distances, i.e., between transistors on a chip or between chips in a single-chip module, electrical interconnects still have headroom for coping with increased requirements. The preferred implementation of optical interconnects inside a satellite depends on the timing of the introduction. In the short term, we expectfiber-based solutions to become the choice of satellite designers, moving towards embedded solutions using waveguides on the PCB in a later stage.

Optical interconnects in very demanding environments for high-througput data-communication applications are a viable means of overcoming the physical limitations of longer electrical lines on PCB boards. The reliability needs in space compel further research into materials, processes, design, and standards for optical interconnects and packaging solutions. We are now performing a preliminary design of a typical satellite data-communication board using close-to-space-grade optical-interconnect technologies to quantify the bandwidth and power gain compared with a traditional board with electrical signaling.

This work was carried out under contract 21873/08/NL/EM of the European Space Agency's European Space Research and Technology Centre (ESTEC-ESA). The authors thank Nikos Karafolas of ESTEC-ESA for his support and project partners at Atmel and Thales Alenia Space for fruitful discussions. The work was supported by the Federal Office for Scientific, Technical, and Cultural Affairs in the framework of the Belgian InterUniversity Attraction Pole Program Photonic Information Systems, the Fonds Wetenschappelijk Onderzoek (FWO), the Geconcerteerde Onderzoeksactie, the European Commission's Sixth Framework Programme European Network on Micro-Optics, and the Research Council of the Vrije Universiteit Brussel. The FWO provided Christof Debaes and Jürgen Van Erps with postdoctoral research fellowships.

Michael Vervaeke, Christof Debaes, Jürgen Van Erps, Hugo Thienpont
Brussels Photonics Team B-PHOT
Faculty of Engineering Sciences
Vrije Universiteit Brussel (VUB)
Brussels, Belgium

Michael Vervaeke has an electrotechnical engineering degree with a major in photonics and a PhD in photonic modeling and packaging of chip-level interconnects. He has authored more than 58 papers in international conference proceedings in his research field as well as in education, and has authored or co-authored 12 publications in scientific journals.

Christof Debaes received his PhD in applied sciences at the VUB (2003). His PhD thesis, on intra-MCM (multichip module) optical interconnects, was a collaboration with Ginzton Laboratories, Stanford (California). His current research interests include optical interconnects, nonlinearities in silicon waveguides, and photonic networks-on-chips.

Jürgen Van Erps graduated in electrotechnical engineering with a major in photonics at the VUB (2003) and obtained his PhD magna cum laude from the same university in 2008. In 2009, he was a visiting researcher at the Centre for Ultrahigh Bandwidth Devices for Optical Systems at the University of Sydney (Australia). He has authored or co-authored 18 ISI (Institute for Scientific Information)-cited papers and more than 65 contributions to international conference proceedings. He is the co-inventor of one patent.

Hugo Thienpont is a professor of photonics. He chairs the VUB's Applied Physics and Photonics Department and is director of the Brussels Photonics research group B-PHOT, which he built over the years and which today counts some 50 photonics experts.

Mikko Karppinen, Antti Tanskanen, Timo Aalto, Mikko Harjanne
VTT Technical Research Centre of Finland
Espoo, Finland

Mikko Karppinen joined the VTT Technical Research Centre of Finland in 1998, and is currently senior research scientist. In 2001, he was a visiting researcher at the University of Ottawa (Canada). His research focuses on fabrication and hybrid integration of photonics devices for applications, including optical interconnects, sensors, and lighting.

Antti Tanskanen earned his MSc in physics from the University of Helsinki (Finland). From 1997 to 2000 he worked as a radio-frequency design engineer for Nokia. In 2000, he joined VTT, where he works in the Optical Sensors and Modules group.

Timo Aalto has worked at VTT since 1997. His main research focus is on silicon photonics. He received his MSc and DSc in optoelectronics technology from the Helsinki University of Technology in 1998 and 2004, respectively. He is a senior research scientist, presently leading VTT's Photonics and Microfluidics team.

Mikko Harjanne received his MSc in optoelectronics from the Helsinki University of Technology (2003). The same year, he joined VTT's Photonics and Microfuidics team, where his work involves designing silicon microphotonics components. He is the author of 18 papers published in international journals.