Hi Mike,
VoIP isn't enjoying the cakewalk in 2001 that some had envisaged six years ago, but it is making inroads in some very subtle, and in some cases, profound, ways that will lead to its greater success, for better or worse, with integration happening over time.
This is not to suggest that its success will ris to sudden ubiquity, and it certainly will not be sudden, even when the ILECs do get around to offering it to residences. It probably will not even be noticeable as an event, or series of events, because the requisite changes will be distributed over longer time frames than those which human attention spans can correlate.
There is no flash cut that's going to happen to make it so. Or even outwardly distinguishable traits, in many regards, because VoIP is not going in as a revolutionary technology by most SP adopters, except in those instances by those who are doing things in niche, or vertical, point solutions. Rather, it is becoming integrated through numerous compromises to its original design philosophy, as conceived by end-to-end IP Telephony advocates, in ways that suggest emulation of what already exists, rather than going in as an overlay of radical new tech.
To examine what is happening here requires that we distance ourselves from proprietary implementations, and even standards-based initiatives for a moment, and examine where the technology (in all of its many forms) "is" and where it "is not" penetrating the marketplace at this time.
It is being used in the interoffice transport spans by a growing number of international carriers, between their international gateway offices. Interstates, too, are adopting VoIP, but in the IXC space (with a couple notable exceptions, such as L3) to a lesser extent. In the interstate space packet voice often means ATM, not IP.
It is also making inroads in softswitch applications, which is what L3 is doing, replacing Class 5/Class4 switches with NT and Unix servers, effectively.
But in the latter two applications, MOST of the traffic that traverses those platforms is usually converted from POTS to IP at some point through a growing number of gateway types, but the voice product is almost always delivered to the end user over a standard telephone, the POTS instrument of your 100 year-plus old ILEC. But even here, in commercial offerings, this is about to change through the use of enterprise IP phones (see below).
Then there is packet cable, which purports to be able to deliver voice services intra- and inter-MSO without ever being contaminated by legacy voice architectures, until it has to hop off-net to reach someone who doesn't have the packet cable service, which means virtually everyone. Here again, the gateway is used.
In the enterprise, where PBXes have run their depreciation cycles and Centrex contracts have run their term, IP Phones are now beginning to surface in substantial numbers. Here, telephone instruments are actually Ethernet LAN end points, just like PCs (in fact they even use the same Ethernet switches in the closets), or they may be soft IP phones on the Ethernet, which are integrated on-screen PC applications, using a phone or other I/O interface that is attached to the PC's bus.
On-net traffic in these instances is intended to be all IP except for branch locations who are still using the PSTN, and Long Distance services which use the PSTN. Interoffice enterprise trunks, however, cold be voip end to end.
But the ILEC's voice trunks that feed desktop IP installation (IP PBXes, or un-PBXes) are still your legacy T1 DID/DOD (direct inward dialing/direct outward dialing) types of trunks, which must once again use VoIP gateway functionality (that may exist in any one of a growing number of elements called by other names, depending on which vedor you use) in order to connect to the new environment.
Some wireless voice architectures have also already begun to use IP for reasons other than for compression benefits. Actually, the compression algorithms used for VoIP and those used for cellular voice are not all that much different in their b-w consumption and effect. In wireless voice, however, the IP component is used at this time for directory purposes and to aid in the delivery of Internet-type services (email, streaming audio, etc.), and eventually for end point identification and discovery purposes as well.
In recent months WSPs in areas of high usage around the globe have been making a lot (A LOT) of noise for the need to implement IPv6, because they have run out of ways to use the paltry IP address spaces that they have been allotted to them. [Someone told me recently that one of the west coast universities has more v4 address space assigned to them than all of China.]
IPv6 would expand the number of end user addresses, exponentially, they argue, while doing away with many of the kludges that now exist in DHCP (dynamic host configuration protocols) and NAT (network address translations). Hong Kong, Korea and Sweden come to mind here, but I'm not sure of which operators, specifically, they are.
If I continue, additional apps become more niche in nature. But you can see what's happening here. Like fiber creep, there is no IP voice in all of its many variations, as more and more applications are adapting to it in the voice realm, leaving only the end office, eventually, as the last hold out in the puzzle.
Which is not to say that both NT and LU and their European competitors have not made provisions for conversion yet, because they have. It's the carriers who have yet to adopt those feature capabilities, and for some very good reasons.
A primary reason being: what end users have in their homes and offices: POTS, PBXes and Centrex services that are still on the books from anywhere from six months ago (when IP Phone releases by the big vendors hadn't even been out yet) to seven years ago. Creep, by definition, takes time.
As PBX and Centrex contracts come to their expiration dates, more and more enterprises are opting for IP phone solutions, if not ubiquitously at this time, then as back fill.
And the encroachment onto the ILECs turf continues. But just as the ILECs are likely to be the most influential players in FTTH via PONs, eventually, they, too, will come around to offering IP voice. If not as a staple at first, then as an option. And they will do this as they have done other migrations in the past, such as when they migrated to T1 signaling instead of metallic relay dial pulsing for businesses. And as they evolved from rotary dialing to touch tone for residences. They will evolve into it, profitably.
I've only touched on legacy landscape and how "it" will be changed. There are other personal level applications that will be fostered by SIP (session initiation protocol) that are closer to the user that will also evolve, and take shape over time. See Jeff Pulver's site pulver.com for more discussions on those. In fact, I think I'll ask Jeff via PM to critique my comments, above, and add some verbiage to this last paragraph, as well.
Jeff? Would yo oblige us here? TIA.
FAC |
