Hi Darren,
You've presented some information that needs a good deal of qualification, IMO. The overstated throughput rates that I mentioned the other day for satellite systems comes into play here, as well. You stated:
"The Teledesic system is designed to support millions of simultaneous users. Users will have two-way connections that provide 64 Mbps on the downlink and 2 Mbps on the uplink."
[What follows assumes that Teledesic gets beyond it's current crisis-of-spirit stage, and actually comes to fruition without an umbilical cord wrapped around its neck.]
This is far too vague of a statement to be making concerning throughput expectations at this late date when we know better. We learn through hindsight, after reviewing the original claims that were made concerning HFC systems and how it is actually working out when it takes up users, that this will only be true during the early adoption stages.
We've come to learn the harsh realities that set in once users take up services in large numbers.
A fairer representation of throughput expectations would be made possible using a sliding scale model depicting number of users, average hit rates per unit time, average file/payload sizes, holding times, packet modality... whether it was mainly TCP or UDP, and a whole slew of other factors which have to do with terrestrial considerations (land-based portions of users links which are indeed a part of end to end measurements), etc. On top of the final figure, add another 10 to 20% for retransmissions due to errors and packet loss during periods of sustained congestion on all systems concerned.
Otherwise, your statement would lead the uninitiated to conclude that they will have access to super-T3 transmission rates on the downlink all to themselves, on an on-demand basis. I think that I once saw such a utilization graph showing some of the parameters I've listed here, but I cannot find it at this time. Do you have such data?
I'd also be a little wary of the latency numbers Teledesic cites, for they too seem to be best case, avoiding many of the delay intensive applications and media types of the future which will constitute the land-based portions of end to end links, such as IP telephony, pcs, cellular, tandem wireless/satellite, etc. FWIW.
Regards, Frank Coluccio |
