The frequency-dependent arrival time [see below] of these fast radio Bursts (FRBs) indicates a delay by a volume of electrons greater than the Milky Way can account for, and suggests that they originate much farther away -— from the universe at large. Indeed, some of them have been localized to individual galaxies at cosmological distances from us.

If the FRB 121102 source is not a neutron star, then there is another interesting possibility: its companion could be a sunlike star. An orbital period of 157 days around the sun would mimic a planet with an orbital radius that is halfway between Venus and Mercury. If the FRB source emits a pair of beams in opposite directions, 157 days would correspond to half of the orbital period —- in which case the orbital radius would be similar to that of the Earth around the sun. This is an intriguing regime, consistent with the FRB signal originating from a transmitter produced by a technological civilization based on a planet in the habitable zone around a sunlike star.

Most stars are much less massive than the sun. An example is our nearest neighbor, Proxima Centauri, a dwarf star that hosts a habitable planet, Proxima b, with an orbital period of 11 days. Interestingly, the orbital period characteristic of the habitable zone around a dwarf star is not very different from the 16-day period exhibited by FRB 180916.J10158+65. Many such planets are “tidally locked”: one side is permanently facing the star, the other facing out into space. Once in each orbital period, this planet would show us its permanent dayside -— the side where stellar power could be harvested by photovoltaic cells.
- Avi Loeb; June 24, 2020, ScientificAmerican.com

By way of explanation for those who are not radio astronomers, the "frequency-dependent arrival time" mentioned in the first sentence goes by the technical name "dispersion measure" (DM). The basic idea here is that when any electromagnetic radiation, like radio waves, passes through a medium containing free electrons, the speed of light is not that of light in a vacuum but less by an amount that depends on both the number of electrons along the waves' path and the frequency (or wavelength) of the radiation.

In short, the electric charge of the electrons acts something like a drag on the radio waves, slowing them down. And since the longer wavelength (lower frequency) waves are less energetic they are effected and slowed more than waves of higher frequency (shorter wavelength). Thus, the difference in arrival time at different frequncies from a short burst of radiation, presumed to be emitted at a range of frequencies simultaneously at the source -- the dispersion measure -- is a measure of the number of free electrons between the source and the observer (i.e., us).

Dispersion measures for pulsars have been used for a long time to map the free electron distribution in the interstellar medium of the Milky Way -- its ionization, since the electrons mainly come from the photo-dissociation of individual atoms by hard ultraviolet light emitted by very hot stars (as well as from cosmic rays). After doing such a mapping, the process can be turned around and used to roughly determine the distance to a given pulsar based on its location (direction from us) and its dispersion measure. The math for this is laid out succinctly at this wikipedia page, for those who might be interested. Note that the dependency is actually on the frequency squared, which magnifies the effect, and as well that the dispersion constant, k_DM also depends on the square of the electron's charge.

Radio astronomers like the dispersion measure because it's an easily made measurement: one simply determines the arrival time difference at two (or more) different frequencies for some blip or bump in the time-varying radio brightness of the source. Again, there's an implicit assumption that the blip occurs at all frequencies at the source simultaneously. For pulsars this is true for all reasonable models of how they emit their pulses. But if the source emits its burst at different times at different frequencies then the dispersion measure can be misleading since it contains an intrinsic component which is difficult to separate from the travel path component. In other words, its unknown how much of the large DMs measured for many FRBs are produced in the source itself.