Hello Steven, (this time I double checked...)
Thanks for the clarification and for expanding on your interpretation of the article. The one point which I want to make is that simply going with an IP approach to voice does not in and of itself yield stupidity, in the sense of dumbing down the model.
Many times, it actually adds to the cumulative amount of network smarts which are actually required, because many approaches add another layer of IP constructs which simply emulate former PSTN ones, due to the continuing need for the PSTN (in those situations I'm referring to, which make up the majority of VoIP types today) at the end points, and even in overflow situations on the greater WAN, anyway.
So, one actually winds up with two smart approaches working in tandem, or interworking, as it were, instead of one dumb one. The end result is opposite, however, to one which would be expected in the way of a Gestalten kind of resolution. Just the opposite, in fact, for the time being, but it is being greatly facilitated and permitted to gestate by regulatory forbearance, in many ways. But that's as far as I'll go on the topic of regs, for the purposes of this post in today's discussion.
The soft switch approach, in my opinion, reflects this very kind of dichotomy of network intellect, due to its continued dependence on gatekeepers and control points, but it is a necessary evolutionary step, or another one like it would be, on the road to a truer form of Internet Telephony which would be solely dependent on IETF RFCs, avoiding the older PSTN rules, entirely. The chasm between here and there is just too great a jump to take at one time, for reasons having to do with embedded infrastructure issues in a massively pluralistic setting such as the global switched telephone network, or GSTN.
Getting back to soft switch, such a step is a necessary and time consuming one to take, however, due to many legacy components, both administrative and physical, which are still crucial for an any-to-any connectivity scheme using available end point equipment, namely the black desktop- and kitchen wall phones, which themselves are among the dumbest things ever invented in networking. Hence, they become one of the collective objects in the stupid network regime's argument. Their stated principle is to replace these dumb end points with smarter terminus equipment, while reassigning stupidity to the network core.
----
Just a word on the jargon now being adopted by edgists, if I may. Have you noticed how the larger players are now touting the ability to carry your data end to end?
Making distinctions between the edge and the core these days is not as straightforward as it once was, when large backbone providers' mesh topologies previously depended entirely on transiting over common core boxes deep in the 'net through peering and other forms of resource sharing.
Today, a large provider may be able to circumvent what has been traditionally known as the core, almost entirely, through various means employing very fat pipes (both real and virtual) between their own locations, or POPs. Although, they will still share paths over core routes they don't control or where they don't have dedicated paths to.
For example, when a provider with the ubiquity of a UUnet, say, can do this, they couldeffectively assign new roles to their edge devices which were previously assumed by core functions, or obviating the core functions entirely, for most purposes. Such would be facilitated by a Juniper type of routing deployment, and by others who are catching on to this newer approach. High capacity pipes between their "edge devices" in a large number of cities and locales make this all possible.
We see this not only on the Internet, but also in large corporate enterprise designs, where backbones are collapsed, or neutralized, through the use of heavy duty switching and routing elements capable of very high processing speeds and port sizes.
Sometimes this distinction is overlooked, and pundits continue to use the old definitions in their new arguments. And so it goes...
Comments welcome.
Regards, Frank Coluccio |
