To: Kenneth E. Phillipps who wrote (652 ) 5/30/1998 12:39:00 AM From: Frank A. Coluccio Read Replies (2) | Respond to of 3178
Welcome Back, Ken. The article was comprehensive in that it wrapped up to the present what has been going on. While I enjoyed reading it, in parts the author sounds a bit like a bandwagon hopper, making claims that are only half thought out. I have to take minor exception to two paragraphs that are overly simplistic and misleading, IMO. ======== "Data vendors have a major advantage in that they must modify only a small portion of a system: They simply convert voice to data, then deal with it as usual. The voice portion of a network becomes just another port into the network." AND "For voice vendors, gearing up for VoIP is another matter entirely. The voice portion of the new network is very small: Quality of service, routing and integration are all done in the data world." ============== For the data vendors he states, "The voice portion of a network becomes just another port into the network. <?> Nothing about inordinate latencies to overcome, the need to optimize packet lengths and timers? Cross-matching codecs using different algorithms? The HUGE problems that will become clear with regard to authentication and privacy matters? The quarantine-like separations between SS7- compliant and non-compliant VoIP nets? The complexities of coordinating multi-vendor gatekeeping architectures, so they remain interoperable? I could go on, but I think you see my point. On the voice side he states, "Quality of service, routing and integration are all done in the data world." Oh, I didn't know that. Guess the voice network quality should come up to the quality and consistency that the Internet has maintained all these years. Seriously, though, is it any wonder that the first predispositions demonstrated by the bell heads was to elect to use ATM and digital communications mux equipment (DCME) schemes, and before that CLEP and ADPCM, all of which were deterministic , and _not_ just "best effort," before they saw the IP writing on the wall? That aside, the author totally ignores some of the human behavioral aspects that voice architectures have been tailored to accommodate for so many decades, and he attempts to reduce the world of voice versus data to a comparison of protocols and primitives. Where the author places his entire emphasis in the voice sector on transport, the major nut that really has to be cracked should be place on voice account and enterprise end-user administration. Not Transport, alone, or even primarily. There have been 16-to-1 economical codec algorithms to handle low-bit- rate-voice, in a deterministic manner, for over a decade. Some of them are right over the horizon, waiting to re-emerge in the Integrated ATM/IP realm, and will integrate with IP very nicely, when they are ripe. Another thing that the author fails to take into account is the cross-pollination that has rapidly occurred in both camps. See another post on this topic at:Message 4402157 Just some of my views. Regards, Frank Coluccio