On Wed, 10 Aug 1994, Dave Emery wrote:
Dave Emery
Your understanding of how IRIDIUM(r) will work is incorrect. It most certainly WILL be the NORMAL operating mode for a subscriber unit (cell phone, if you will) to talk to another subscriber unit by only going through satellite links. The caller will be authenticated via a "home" equivalent to the MTSO switch, but the call itself will NOT go through the switch (or any other) unless it is to a phone number which is not a subscriber unit. ONLY in that case will the call be routed through the MTSO equivalent.
Thanks for the correction - there is not a lot published about the system that I'm aware of (at least in technical journals I see) so I'm apparently out of date on how the current system works. But your qualification about going to a phone number which is not a subscriber is a very big one. No doubt IRIDIUM service will cost more per minute than some current ripoff prime time AMPS cellular costs and even perhaps in the outrageous INMARSAT ($>6.00 minute) range and is unlikely to replace all but a small fraction of current wired phones and terrestrial cell phones, let alone the hordes of PCS and cable company phone connections coming in the near future. So on a statistical basis an IRIDIUM subscriber is rather unlikely to be calling another IRIDIUM subscriber. I will grant you that if IRIDIUM becomes competitive in remote areas that a certain amount of remote area to nearby remote area traffic will be IRIDIUM transported, but my guess is that nevertheless most IRIDIUM traffic will be to numbers outside the system and thus go via the MTSO equivalent. This does raise the point, however, about what the IRIDIUM system plans to do about pirates who wait for an IRIDIUM to IRIDIUM call to set up and then take over the uplink with higher power (probably just using high gain steerable antennas would do this fine) and talk on someone else's nickel. I imagine that if the satellite actually demodulates the digital voice/data stream to baseband and switches it as digital data rather than rf that it would be possible to incorperate cryptographic authentication of the packets and have the satellite borne switch check all its packet streams for valid user id. But of course this adds a weight and power penalty to the satellites... Do you know if this problem been thought of and addressed or is it being assumed to be as impossible as AMPS cellular spoofing apparently seemed to be to the developers of that system ?
Your thoughts about caller authentication are correct. I don't know if IRIDIUM is planning to do this correctly or not.
It had better. Dave Emery