WW -
1. Please comment on the fact that while CDMA originally promised a 10x capacity improvement, actual field results show that capacity gains on a system-wide basis have only averaged 4x - exactly what non-CDMA "proponent" mathematics predicted.
I'd love to see a reliable survey that showed CDMA performance vs analog. However, 4X is extremely non-credible, although I've seen biased reports which said similar things. For instance, there was the paper which said that if you assumed a constant bit rate (i.e. constant talking), then CDMA got about the same perfomrance as GSM. Well, no duh! (there are many ways to slice and dice this, but undoubtedly one benefit of CDMA is its stat mux capability) Or there is my text book which compares the performance of a single cell (with constant bit rate) of CDMA with a single cell of TDMA and CDMA comes out the loser. Again, no duh! Sadly this kind of GIGO is all too common in technology wars, and often it isn't even conscious.
2. Please comment on the fact that CDMA equipment costs 3x more than the analog equipment it replaced - resulting in a relatively small improvement in capx cost per erlang for the carrier community which deployed CDMA. Juxtapose that metric against those carriers which deployed other digital air interfaces.
Undoubtedly true, but not entirely relevant. On a erlang basis the equipment cost for CDMA is more than for TDMA. But equipment costs are ever shrinking proportion of total costs, (i.e. other costs such as cell splitting, real estate costs, site taxes, ... are an ever greater portion of the total cost). Thus, even if a technology costs twice as much for equipment cost per erlang and always will, if both systems equipment costs are shrinking then eventually the more spectrally efficient systems win out.
. Please comment on the system design issues associated with the deployment of a high speed mobile wireless access system using QCOM's version of CDMA. For example, please discuss the implication for the system designer of the varying Eb/Nt requirements for the multiple modes of HDR. Here's a hint.
Thanks for that data (no sarcasm intended). And I agree that the protocol for HDR must be pretty interesting. But TDMA is even more interference limited and thus the protocols must be very significantly more complex. BTW, the most interesting thing about the tables is that the penalty for having one 154K user vs two 77K user is about what you would naively expect (3dB), and the doubling effect remains as 'expected' until you get beyond 1.23M (i.e. beyond QPSK). At that point the added power needed to double the bandwidth goes up to more than 6dB. Without more data it is impossible to determine exactly what this means, but naively it appears that only very rarely will a user use the max data rate. It is just very power inefficient (and hence spectrally - due to interference with other cells).
All in the interest of debate and information flow.
Clark |
