Sign In

Communications of the ACM

ACM News

Terabit/s Optical Data Link Demonstrated for Satellites

View as: Print Mobile App Share:
Satellite data links.

Substantially increasing the data transmission capacity of satellites will make Internet access cheaper and easier in remote areas not connected by optical fiber to the terrestrial Internet.

Credit: Hira Kalid and Saima Zafar

Data transport via glass fiber has become the backbone of the terrestial Internet. Such optical fibers now carry the bulk of all data — thousands of terabits per second — over large distances. Wireless data transmission still depends on radiofrequency (RF) datalinks, which typically have a transmission capacity of a few gigabits per second (gigabit/s), and their capacity cannot be scaled up beyond 100 gigabit/s.

This is about to change, as a field experiment in August demonstrated. A wireless optical data link was established between two locations in the Netherlands: a ground station in Lopik and a transmission tower in Ijsselstein, 10 kilometers to the east. On that day, the data link's capacity was just 28 gigabit/s, but according to Wimar Klop, systems engineer of the TOmCAT (Terabit Optical Communication Adaptive Terminal) project, scaling that up to 1 terabit/s is straightforward.

Substantially increasing the data transmission capacity of satellites will make Internet access cheaper and easier in remote areas not connected by optical fiber to the terrestrial Internet. It would also enable broadband Internet connnection on passenger airplanes and ships. Also, terabit/s optical data links via satellite would present an alternative to transoceanic data cables that are quite vulnerable to sabotage in times of international conflict, as demonstrated by the recent demolition of the Nord Stream pipeline by unknown actors.

While the components of this technique had already been demonstrated separately, either in the lab or in field tests, this was the first comprehensive field trial in which an RF (radio frequency) signal was converted to the optical range and back to RF, while transmitting data with advanced noise reduction and error correction. The results were presented in October at the annual International Conference on Space Optics (ICSO).

The wavelength the lasers used was in the optical range, but not visible to humans; it was in the near-infrared at 1550 nanometers, so its frequency was 190 teraherz (for all electromagnetic waves, including RF and optical, the product of frequency and wavelength is the speed of light, c). Data transmission capacity is fundamentally limited by the frequency of the carrier signal. In the case of RF datalinks, that frequency is 2.4 or 5 gigahertz, so thousands of times less than the frequency of an optical carrier signal.

The fact that the infrared laser signal was transmitted virtually error-free more than 10 kilometers at low altitude means it can also be transmitted through the atmosphere all the way up into space, to geostationary satellites at 36,000 kilometers (22,300 miles) altitude. The reason is that the density of air decreases rapidly with increases in altitude, so the thickness of the entire atmosphere is equivalent to that of about seven kilometers of air at sea level. (Once in the vacuum of space, the signal does not degrade.) The laser light cone, a few centimeters in diameter when it leaves the transmitter, spreads out to a diameter of 800 meters by the time it reaches the altitude of geostationary orbit.

Optical datalinks between satellites – in a vacuum— have been in operation for several years; the SpaceX Starlink satellite Internet constellation uses such a system. The difficulty is sending optical signals up through the turbulence of the atmosphere to reach satellite altitudes intact.

Adaptive optics, originally developed for astronomy, tackles this problem. The basic idea is to reverse the deformation of the laser wave front carrying the signal in advance. When a pre-deformed laser wave front hits atmospheric turbulence, pre-deformation and deformation cancel out, delivering a near-perfect laser signal at the receiving end.

Of course, this requires knowledge of how exactly the turbulent air deforms the wave front. This can be deduced by sending seed laser signals between the transmitting and receiving sides of the optical link. The extent and manner in which the seed signal is deformed is then translated into a set of instructions to actuators, which deform a small, flexible mirror at the laser source that pre-deforms the data signal before sending it to the receiving end of the data link. As turbulent air changes direction and speed all the time, the actuators must be adjusted, sometimes as frequently as 5,000 times a second.

The seed laser only requires milliwatts of power, while the actual data signal requires 50 watts per 100 gigabitb/s channel. In the future, data links with multiple channels may beam up 500-1000 watts of narrowly focused infrared power. Says Klop, "It will be necessary to install safety systems that detect aircraft passing overhead."

Even that much power will not be able to penetrate heavy cloud cover, which is a disadvantage compared to RF datalinks. According to Klop, "Plans for datalinks to geostationary satellites involve multiple ground stations in southern Europe, where the sky is mostly clear. Statistically, it will be possible to achieve availability of at least one datalink 99.9% of the time."

In a related development, in late September TNO, the Netherlands Organization for Applied Scientific Research, delivered its first SmallCAT laser communication terminal for testing to Canada's UTIAS Space Flight Laboratory. SmallCAT will be on board a Norwegian satellite scheduled for launch in 2023, which will establish an optical datalink to Earth's surface.

Optical datalinks have other advantages as well. RF signals cannot be focused very well over large distances, which makes them prone to eavesdropping. Also, multiple RF signals crossing paths will interfere with each other, which is why RF datalinks in many cases require licenses, and scarce bandwidth has to be bought from governments. No such restrictions apply to optical and infrared datalinks.

It is not expected, however, that optical datalinks will completely replace RF data links. The latter will remain the standard technology for small- to medium-capacity wireless links, while optical datalinks will be put next to large terrestial Internet hubs or datacenters to provide the communication backbone to satellites. Also, its resistance to eavesdropping is of special interest for military applications.


Arnout Jaspers is a freelance science writer based in Leiden, the Netherlands.


No entries found