How to connect Earth and Mars with an internet link? Here is a quick explanation of my Marslink simulation (v3). Design your own interplanetary network and find out the total Mbps, latency and cost! I'm pretty sure SpaceX will deploy a version of this one day.

Design your own at It's a free website not collecting any data. Note that a computer screen is needed due to the number of UI elements to display - not optimized for smartphones.

The Deployment Report module calculates details of required launches and costs for your exact design.

You can also see the detail of individual rings and their launch 'schedule', including tankers.

There are some approximations in the calcs but I think it's a good start! Let me know if you have any questions or suggestions. I tried to do a very short video but it was still 5 minutes and couldn't get into any interesting detail. I'm adding a video on flow calculations.

Flow calculation explanation: A planet's high data throuhput is like large river. This river splits into smaller, low-data-rate streams. These streams later merge back into large, high-data-rate rivers, reaching the other planet. High rate is depicted in red, low rate in blue.

In case you want to access the previous versions, they are available at and

@SpaceX Could someone explain how you can get over 200 Mbps speed when your latency is over 15 minutes?

Yes. A laser link can carry a certain amount of data per time or per distance. This amount depends on the wavelength, the modulation, the superposition of signals, etc... So let's assume an individual laser can carry 100Mb every second that it's transmitting. If the receiver is 10 second away (at the speed of light), then the latency is 10 seconds. It takes 10 seconds for the information to reach the receiver, and it receives 100Mb per second. Also, note that 1000Mb is in space at any single time! Now, let's assume the distance is much greater. There are a bunch of relays to avoid signal loss. If there is a relay every 60 seconds, you need 15 relays. They all maintain 100Mbps and the total travel time is 15 minutes. Because the position of planets relative to one another changes all the time, the latency changes constantly. The network topology changes also based on distances between satellites, and there are multiple paths, so the latency is actually a set of latencies.

@SpaceX I see, thank you for the explanation... but this does not mean two-way communication at 100 Mbps... it just means one way. Which, I guess, makes sense. I was thinking more in terms of a video call... where the latency goes both ways in real time.

This would work both ways actually, as a transmitting laser would be collocated with a receiver. So if there is 100mbps one way, there is also 100mbps the other way. Phone calls are out of the question due to latency. But you can send a video saying hi and talking about your day. This is how Mars comms will look like I think.

@AAristi1976 @SpaceX Hi @julienvm : "Phone calls are out of the question due to latency" - I actually invented a system with No latency. It is based on the latest advances in Quantum physics. Ready to listen? 👋

@AAristi1976 @SpaceX Ok, shoot!

@SpaceX I really like your voice, you have a deep soothing voice

@SpaceX Haha thanks bro!

@SpaceX This is unnecessary complexity, kinda against SpaceX's entire company design philosophy. 3 or 5 large starlink relays in a lower orbit, closer to the sun than Venus. 3 or 5 transceivers in polar orbit around earth and Mars, and an array similar to earths.

@SpaceX This is simply a simulation reflecting how physics work. You can set your own assumptions as inputs for the laser link capabilities. I haven't presented what I would consider a realistic design. I think SpaceX will want at least 1Gbps.

@SpaceX Also a more realistic approach would be how Google and Netflix set up their data in many small servers at ISP locations. Instead of all requests going to a few large data farms, most content is stored closer to the end user.

@SpaceX We can try whatever we want, the speed of light is and always will be the limiting factor.

@SpaceX Yes, until we can use particle intrication!

@SpaceX Quantum entanglement cannot transmit faster than light.

@dm_jns @SpaceX However it can't be used to transmit information because the measurement gives a particle a random state among the possible options. So you can't force a specific state.

@footycomps2024 @SpaceX Exactly.

It's so cool, that you have built the sim I also planned.👏😊 I wanted to run genetic algo optimization for choosing the best config for the satellite set to reach specific target performance. Other: did you build in half-statits in the simulation too? Those could orbit the Sun at any desired slower velocity - using solar sail to counteract the proper portion of Sun's gravity. I will check the sim and come back with some more remarks. 👍

@SpaceX great! it is on github so you can branch / contribute too

@SpaceX Cool. I planned a seasonal energy storage sim next, but the temptation to check your program seems strong. 👍

@SpaceX Starship's diameter is 9m but telescope max diameter in your simulation is only 5m. I think I would prefer 9m diameter telescopes and greatly increased transmission power to reduce the number of satellites needed.

@SpaceX I think it would be tough to mass manufacture and then stack 9m telescopes in the rocket honestly. You need 4 laser ports per satellite. 5m already seems impossibly large.

@SpaceX Why 4? Isn't 2 enough? Just give each satellite 2x 9m dishes pointed in oppsite directions with a gimbal between. Launch one satellite per starship.

@SpaceX I think you need 3 ports for an eccentric rings configuration (what you seem to have in mind): 2 for in-ring links and 1 to communicate outside the ring. However I think circular rings are best, and those require 4 ports.

@SpaceX No, the satellites can be in a circular ring and they don't have to send the signals all the way around the sun. LEO-satellite satellite-satellite x N satellite-LMO No need to send signals to the satellites that are not part of the signal chain between Earth and Mars.

@SpaceX So you assume zero redundancy I'm not against your idea but I'm now sure it's easier than many satellites.

@SpaceX Very kewl!! Doesn't this sort of concentric ring, many many hops between rings, design hugely increase end to end latency?