Terry,
The author of the article which you referenced in Post # 45 was, in at least one instance, grossly ill-prepared to write about the metrics of this topic, and the overall impact of this technology. I feel that I must correct one aspect of the message, just to keep things in their proper perspective.
The author states (quite innocently, I would presume):
" Just think, you will be able to increase the capacity of a single
fiber by up to 128 times using a single chip. For example you will be
able to carry 128 phone conversations at the same time instead of just one without, any interruption."
Ouch. To highlight how far off his message is from reality, 128 wavelengths, called lambdas, each supporting 192 individual T3s (in an OC-192 stream), can support 16,515,072 'worst case' calls, each at the standard 64 kb/s rate. This is the telco default rate, determined by legacy channel parameters.
128 lambdas/strand * 192 T3's/lambda * 28 T1's/T3 * 24 voice/T1
= 16,515,072 voice calls.
Mildly compressed to 32 kb/s (which is very common on the public switched telephone network), the figure jumps to
33,030,144 simultaneous voice calls.
If cellular-like compression protocols are used at ~ 16kb/s ea., including overhead, also now common in many parts of the network, the new number becomes
66,060,288 calls.
When straight voice-over-Internet-protocol, or VoIP, is used, which is where the future is taking us in the next few years, it doubles again to
132,120,576 calls.
With enhanced VoIP traffic management, making full use of available pause times between talkers' syllables, this figure again, at a minimum, is capable of doubling to
264,241,152 calls, all over a single strand.
The number doubles again, with higher compression factors, to over a half billion calls, if one could ever find enough people (a neat billion, in order to account for one talker on each end) to fill such a pipe.
If you double the electronic rate from OC-192 to OC-384, or if you double up on lambdas, take your pick, the figure jumps to over a Billion simultaneous calls.
Using a realistic figure for the sake of discussion, 264,241,152 calls is a far cry from the 128 uninterrupted simultaneous calls which he stated.
Consider the amount of fiber being laid, and then think about the capacity requirments of both voice and data, using the above figures as a backdrop. These figures should serve to demonstrate the validity of the current claims being made that voice will amount to but a squirt in the larger scheme of things, going forward.
Getting back to the article, I feel rather certain that the author meant well. I just couldn't permit it to stand that way, unqualified, as it were. FWIW.
Best Regards, Frank Coluccio |
