To: ahhaha who wrote (458 ) 11/24/1999 11:48:00 PM From: Frank A. Coluccio Read Replies (2) | Respond to of 1782
re: >"I got it I guess." And I guess that I don't got it. And I am guilty once again of assuming too. Apologies. Let me see if I can find a web-based tutorial with a good set of graphics on basic telephone switching, so that we can have a common point of reference for discussion. I'll post it when I find one that's suitable. In the meantime, if anyone else here knows of one, please post it. The ones I have in mind have too much Intelligent Networking (IN) and SS7 to be useful at a fundamental level. I think that a part of the confusion here is that you are assigning sharing attributes to the POTS realm, when in fact POTS demands "dedicated" links for each call. This is understandable, since almost all other forms of media do share: The Internet "cloud," CableModem (over a single black cable or HFC), many wireless schemes, etc. In contrast, and as a good case in point, note that each telephone subscriber has their own loop all the way to the central office. At the CO they have their own line port on the "end office" switch. And when they connect to a tandem switch, or to the remote switch, they use a "dedicated channel" to get there. Once at the remote switch location, they connect to another dedicated loop to the called party's location. At no point along the way is there any sharing taking place at the individual session level. Where sharing "does" take place, it's in other operations related elements such as billing, surveillance, route pointing, etc., but not in the actual "talk" path. Talk paths are home free. Since each call in POTS is set up on a dedicated set of facilities for that call alone, those calls possess what is known as a deterministic quality because they don't have to share or contend with any other calls. This, as opposed the 'net's sharing approach, which is not so deterministic (far from it, often). Sharing on the 'net, or even in LANS, is most often accompanied by "gueuing" and algorithm-based "arbitration" and now prioritization schemes, which most often relegates traffic to excessive delays when the queues are extensive, jitter, and dropped data. In POTS, these are not a concern. Instead, the higher ultimate costs associated with having dedicated links is the primary concern. Or at least, that "was" the concern. Today, dedicatedness also leads to other pitfalls, such as strandedness, in the sense of ubiquitous reach. IP enjoys the latter, whereas POTS doesn't. Tradeoffs. Hence, Internet based sessions require special QoS treatment because users are in constant contention with one another due to the sharing process. POTS does not require this special treatment. POTS, by definition, has enjoyed "QoS" all along, in its native form. Let me see what I can find, and I shall return. Frank