Hi George, in paraphrasing me, you noted:
"... there's no MUX beyond what is done with a single wire, hauling millimeter wave analog frequencies to the modulator."
You're referring to what I said about the "in" direction: the lower order tributary side of the node, looking back in towards both the subscribers (assuming aggregation is being done in an end-office) and other networked nodes (assuming a mesh), or any combination thereof.
This would be the case regardless of whose super-engine is used. A disjunction occurs, a divide, so to speak, between all of the legacy forms of tributary creation (SONET, TDM) and the newer optical methods of combining streams, regardless of whose approach we're talking about, unless we are discussing another graduation of SONET in TDM (for example the creation of an OC-384 or OC-768, etc. per every lambda or individual strand). The latter (OO-768), incidentally, equates to the 40Gb/s mark.
I think that the disconnect from traditional data rates is most pronounced in the SR model, probably because of its approach to optical multiplexing, or superimposition, which, to date, has not shown any predisposition to following the normal progressions of established hierarchies. Conversely, you'll note that even those networkers who profess IP over SONET, Pysical Optical, and lambda directly, are still packaging their streams in ways that are consistent with SONET/SDH I/O framing formats, albeit in a way that employs a much "thinner overhead." And a very small percentage of emerging box types do not use SONET headers, but these are proprietary and still in trial stages, where wide area deployments are concerned.
SR is protocol- and data-rate- independent, in this regard. Looking "out" towards the line (on the greater WAN) this is actually an advantage n some ways. Looking the other way, "in" (back to the subscribers), however, SR faces a challenge that is less clear to me at this time.
In any event, they would need to either OEM or partner, in order to bridge a significant development effort in these regards, if they themselves are not bought out first. If they pursue a purist approach to both the ins and the outs, then they are in for a very long development ride. And time waits for no one in this game. They would therefore be a buyout mark. Frank_C. |
