Mike:
RE: your request to add 'color' to what a media gateway controller (MGC) is.
Technical:
The MGC is simply a big possibly fault tolerant NEBS compatible rack mounted 48volt
hot card-insertable computer running call control software. It is connected to one or more
managed IP networks with possibly redundant connections.
On one side (the 'subscriber' side) it takes in a voip device management protocol-
basically the signals generated by your phone, but put into IP packets. So it knows when
you go off hook, or press a dial key, or perform some other function at the phone that
signals the network that you want to do something, like get a connection to the network.
So the MGC manages the subscriber side voip media gateway between the subscriber
and the network. This is called the Media Gateway Control Protocol, or MGCP.
On the other side, or the network side there are a number of interfaces. One is to the
PSTN and is an IP version of SS7- it allows you to managed set up, tear down, telephony
services (such as caller id), as well as 800 numbers, credit cards, and all the usual services
expected of the network. There is also an advanced interface that allows one media
gateway to exchange information and control with another, called 'session initiation
protocol', or SIP. This is the fun protocol, in that it enables more internet like services
beyond the existing ones you find in telephony. There are many other mostly unspecified
interfaces for billing, operations, administration, and maintenance.
So all it does is wait for a caller to go off hook, analyze the digits, route the phone call to
the other end party, and bill for it. No problem, right? Why not just replace it with a
browser and give it away for free?
Money:
Ok, enough technical. What the MGC really is, is software, and mostly the same
software you find in the embedded central offices that have been developed since the late
1960's. We have a lot of data on how much effort it takes to build modern call control
systems,
Now it takes about four years and 200 people to write the CO software of the 1990's,
and it will take about the same time and staffing to create a quality MGC's in the 2000's:
the complexity is about the same, and something like 80% of the CO software is
operations and maintenance, which has to retrofit into existing customer care and billing
systems. Just testing it will take a year. You will hear a lot of hype about working
systems, and trial systems, but be skeptical.
And in the end you haven't saved any money for the $160,000,000 in programming
investment you just spent to build a MGC; you probably will also spend about $30-40 M
per year in supporting successfully deployed systems. So you have got to pay for this out
of your hardware margins and software fees for services. At the moment, the DSP chips in
the media gateway are more expensive than the line circuits in the CO, and we haven't
found the killer app in voip yet that will generate new revenue.
My guess is that the real reason Lucent and NT are building the MGCs is the big service
providers are insisting on them. So the only market here is for:
1) LU and NT for replacement of existing equipment with new VOIP based switches.
However, the margins are very bad, I think, and the customers expectations are high- not
a good place to invest
2) Small start up companies. If any have staffing with switching experience and can build
a good product in a short time, they might be a candidate for buyout- I don't know of any
at the moment, but LU or NT or CISCO may have to buy someone, and pay for the
$160M in $1B of stock, or something.
3) Cable equipment companies (NT/ANTC, ADC, Tellabs, and possibly CISCO and
Lucent). The cable providers can bundle voip with multi-media services and generate new
revenues that can subsidize voice. However, cable will likely first deploy using old Central
Office switches, as I've noted in previous posts. Here, NT and thus ANTC seem to be a
likely player. Lucent once backed out of cable voice and is not a player today, but might
return. CISCO has not the strength or success in telco they have in enterprise, though
they would like to get into cable, and they are not call control switch people. ADC and
Tellabls also are not currently in call control.
So the mostly likely market for voip MGCs will be cable in the next generation
deployment, in, say, five years.
justone opinion |
