Look at what the DD-WRT project's mesh applications have done and apply the SAME process models to the aging sat fleet. In one evaluation- much of the older and getting old sat fleet is comparable to a WRT54. In orbit, already ripe for rework.
The correct Solvings are in CONCEPTS rather than specifications. Heavily redundant meshes can neatly handle "Latency Ignored" traffic. The "How to" I propose for affordable builds has an amusingly quaint approach. Recycling! Re-assign existing satellites to turn OUTWARD. And move some to applicable stations along the route .. Yeah- they will need non-trivial work in "some" cases to refit for the new missions. And we need either robotic or human on-orbit work capabilities or this stays fictional.
BUT- the fractional costs compared to build-launch-deploy new sats might sell the idea. Example found in just how many birds we burned up as opposed to deploying for space relays. Bolt them onto other traditionally WASTED transfer vehicles- refuel- and send out to stations for a Deep Space Mesh. Oh- yeah part 2- Fuel comes from...Lunar Fuel Depot or the Martian Processors that Zubrin's group is betting on . Hell, The past wasted ISS resupply craft might have been resources enough to DO this as salvage. The Photovoltaics we've burned up in deorbited birds arguably by now could have made a fracking Powersat- even one directed OUT to power Mesh Elements?
Damned exciting and could be a trifle financially lucrative in a few generations. Imagine having bought International Cable Routes when Morse was the best mode existing.
Teal Deer: Use what we have been raining down on our planet to build a mesh outward bound!