re>>...seldom implemented, to the best of my knowledge, exactly the way it is spelled out by the consortium...
That answers one of the questions I had. So in essence it's more of a politically correct public face to appear aligned strategically, when in fact there is little cohesiveness.
Also, because of its age, I presume there are no provisions for "competitive access" designed in. It would have to be a retro-fit band-aid like it is to all existing networks, true? I suppose the argument would be "it's ATM, it can accommodate anything, therefore there is no issue."
Would you agree that all networks going forward must include carefully thought out architectural constructs to include multiple service provider add/drop capability?
In a sense this leads into your next comment:
"...networks that were conceived by the telco regime...were ONLY designed to target residences...and were never, however, conceived as full service platforms to "all" potential end users"
This would seem to add value to the "classful lambda" approach where wavelength partitioning is by service class. In many (maybe most?) cases, this is synonymous with service "provider", and the classes/providers could be added/dropped at will, and grouped into all permutations of class bundles to any given segments of the network. This would necessarily mean that a) we have lambdas out the wahzoo, and b) we don't care how efficiently the bandwidth within a given lambda is used. So, "lambdas" become the resource, and bandwidth is just a by-product and sort of loses its significance. It's hard to imagine--assuming we could have hundreds of lambdas to end users and thousands at higher tiers--lambda exhaust, but I'm sure it would be made an issue "by design." |
