Geostationary Satellite Interference
As far as I know, there are no satellites in geosynchronous orbits using 60 GHz or 70 GHz signalling, but there may be by the time the first server sky arrays deploy. While server sky arrays will eventually be placed in inclined orbits, the first arrays will be in the same equatorial plane as the geosynchronous satellites, and transmissions intended for one service may be received by the other. This will primarily cause problems for small server sky ground antennas with relatively wide transmit and receive angles.
60 GHz interarray signalling does not penetrate the atmosphere, and is likely to find wide use for space-to-space communications. Server sky arrays will send many beams back and forth, but they will be very narrow, perhaps 100 microradians, spreading to a 6 kilometer spot projected on the GEO orbit at maximum distance, and attenuated by 6x compared to the signal received by another server sky array closer by. The simple solution is for a server sky array to not transmit during the brief times a GEO asset is passing through that spot, and not to listen from that direction. The GEO orbit is 265,000 kilometers in diameter, and rotates at an apparent velocity of 15 kilometers per second; simply interrupt communications for half a second, then resume. If there are two hundred 60 GHz satellites in GEO, this results in a precisely predictable 0.6% interruption in transmission. If the GEO satellites at 6.6 RE forebear transmitting and receiving beams that skip as low as 2 RE , there will be no interference. This may be required anyway, since comsats may not have enough power to send 60GHz high bandwidth beams 13 RE across their orbit.
70 GHz groundlinks to equatorial server sky customers will raise the most difficult interference issues. Small ground antennas will not have nearly the same longitudinal and latitudinal precision as the orbiting arrays. Antennas a few latitude degrees away from the equator will be aiming at arrays with a different elevation than antennas aimed at GEO. But server sky antennas aimed straight up at the equator will see both. If the GEO satellites are listening straight down to the equator, they will hear server sky antennas, too; if the server sky ground antennas are listening to a particular azimuth for a specific array, there may be radio energy in the same band coming from a GEOsat, too. This will probably require bandwidth coordination, and good ephemerides; the server sky arrays in orbit will neither listen to or transmit to ground areas that have GEO ground antennas, and also will avoid the frequency bands a satellite might be using.
Since 70GHz requires 16 times the power to penetrate the atmosphere at 45° elevation, and suffers far more from rain attenuation, it may never be used for non-point-to-point services; the frequency band offers no advantages for direct broadcast television and similar broad-area coverage. Point-to-point requires large aperture antennas for selectivity and spectrum reuse, with very accurate mechanical placement or electronically adaptive control of phased array emitters; difficult and expensive for big iron satellites with COTS silicon. If server sky catches on, it will have so many advantages over GEO orbit big-iron satellites that they will probably not be economically competitive. While shut-out competitors with ITU influence may thwart server sky technology with regulations and denial of licenses, the best way to earn their cooperation is to share the technology with them, and grow the global market as fast as possible.
In time, server sky ground antennas will get larger, evolving to more angular selectivity in both azimuth and elevation. Orbiting arrays will get larger and more selective, and there will be many more of them. At some point, they will "break the inclination barrier", and arrays will be deployed in toroidal orbits with significant north and south inclination. That will increase the potential collision risks, but it will also provide service to more northerly customers, and allow customers directly on the equator to point their antennas away from satellites in GEO. Hopefully, operators will grow more adept at managing collision avoidance and using lightsail manuevering to stay precisely in their assigned orbits.