Path: utzoo!utgpu!watserv1!watmath!att!dptg!ulysses!andante!mit-eddie!media-lab!snorkelwacker!tut.cis.ohio-state.edu!zaphod.mps.ohio-state.edu!sdd.hp.com!ucsd!ucbvax!udel!Mills
From: Mills@udel
Newsgroups: comp.protocols.tcp-ip
Subject: Re:  Mobile TCP/IP (was Re: Can subnets be separated by another net?)
Message-ID: <9007171938.aa18804@huey.udel.edu>
Date: 17 Jul 90 23:38:24 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 32

Phil,

MSS was to put up 240 satellites in 800-km orbits by flinging them from
the Shuttle at odd moments. They were to have steerable antennas and
operate at Etherspeeds. Assuming random (!) orbit inclinations, a ground
station would see about five satellites at a time and a satellite would
see about 35 other satellites at a time. The average crosslink between
the satellites would be about ten minutes.

For homework tonight, you get to design the crosslink acquisition and
routing algorithm. You also get to figure out how a new satellite not
knowing the orbit elements or ephemeris of the other satellites finds
its friends with a 5-degree beamwidth antenna.

There will be a test in the morning. Open book.

Oh, I forgot. The system is power-limited. The best routing may be achieved
with minimum-power routing, rather than minimum-distance. In other words,
take the usual linear metric (like hop count or delay) and square it. Then
construct the minimum path. Surprising things happen, like this can result
in the MAX number of hops, rather than the usual MIN.

I worked on the project for a short while, but have not seen the final report.
However, my simulations showed that the degree of connectivity quickly blows
away SPF algorithms and blows everything away with min-power routing.

Iridium is a rare metal probably ill used to build satellites with. You may
remember the ill-fated XTEN network that could be described as an Ethernet
radio operating at 10 GHz for metro-area coverage. Are we now seeing XTENs
in the sky?

Dave