AT&T and Comcast
Those who can remember back to the last century (1998 to be specific) may recall that shortly after AT&T purchased TCI, they laid out a "three-step plan" for cable telephony. Phase 1 in early to mid 1999 was to deploy circuit switched HFC telephony; Phase 2 starting in late 1999 called for first deployment of packet telephony with local loop bypass bundled with AT&T long distance service; Phase 3 in 2000+ called for end-to-end packet telecom and connecting all LD-IP telephony traffic to the AT&T national packet network. So how is the industry doing with respect to those long-ago goals?
It's little surprise that those overly-aggressive goals were way out of line with the realities that have been achieved. Comcast, which now owns AT&T, has more pressing priorities in 2003 than a full-court rollout of IP telephony. Brian Roberts' opening talk indicated that the 2003 applications focus will be digital cable, cable modems and VOD.
That's not to say that cable telephony is out of the picture. The NCTA reports that as of June 2002, U.S. cable telephony served approximately 2.1 million local voice customers. As of September 30, 2002, AT&T had 1.3 million circuit-switched subscribers and Cox had 651,000.
Comcast had made a conscious decision not to deploy circuit-switched cable telephony, prefering to wait for PacketCable VoIP telephony. Now it needs to figure out how to serve and migrate the AT&T circuit-switched customers, while simultaneously shaking down and starting to roll out PacketCable.
During the week Comcast completed the acquisition of AT&T Broadband, we visited their offices in Philadelphia. Our meeting with Steve Craddock, Senior VP New Media Development, focused on how Comcast's telephony plans are shaping up. 2003 will be a year of limited VoIP residential deployment, with a full rollout in 2004.
"VoIP -- This is No Drill"
Some of what we heard from Steve was later echoed in the session "VoIP -- This is No Drill" he moderated at BroadbandPlus. Cameron Gough of AT&T Broadband (now Comcast) pointed out that the biggest challenges in rolling out VoIP are "integration, interoperability and back office". What Comcast will be doing in 2003 is putting the system into the real world in one specific area and "making the template work right". "VoIP will be the next big thing after 2003."
For cable operators, VoIP telephony means the implementation of PacketCable, which Ted Griggs of Syndeo divided into two eras, BC and AD -- where BC is before CableLabs and AD is after (their) direction. Stan Brovont of Arris pointed out that although there had been talk in the past of cable telephony for second-line service, "primary line, carrier grade" (i.e., life-line service) is now the predominant industry assumption for meeting the bar. The majority of today's cable telephony subscribers are served by Arris equipment for circuit-switched telephony, and Arris is busy trying to work toward maintaining a significant role during the migration to packet switching.
It is interesting to track how PacketCable has developed. The goal was to avoid repeating the telco history with legacy circuit-switching products - these take forever to modify and cost lots once their buyers are locked into a specific vendor's proprietary approach. PacketCable's intent was to build on the increasing adoption and ubiquity of packet technology, leveraging the economies of generic servers providing individual telephony functions such as record keeping, call management, announcements,
Over the past several years, various companies have specialized in creating the multiple separate elements needed for PacketCable telephony. This has created numerous technology options, but put the complex system integration and operations burden on the MSO buyer or a hired integrator. With the entry of vendors like Cedar Point, there seems to be some swing back to integration (this time of the switched IP architecture) in order to simplify the delivery and management of VoIP services.
Judging by the huge attendance at this Broadband Plus session -- all seats were filled and folks lined the walls and sat on the floor -- cable telephony development is very high on the agendas of both the MSOs and the vendors. Responding to an audience question about whether the cable upstream can handle substantial volumes of IP telephony, Craddock said he was confident about using DOCSIS 1.1 today and later moving to 2.0. He also indicated the importance of working on royalty-free codecs, some of which are as good or better than those for which royalties must be paid.
The Impact of SIP Telephony
Another question from the audience was "What about companies like Vonage and Net2Phone as competitors?" In our minds these companies are quite different. Net2Phone is seeking to partner with MSOs: its Cable Technologies Division has developed an outsourced standards-compliant telephony solution for cable operators, using DOCSIS technologies and components from companies such as Arris and Motorola; they are currently in Phase 2 of a trial in Puerto Rico with Liberty Cablevision. Vonage is a competitor: using SIP telephony, it is trying to collect all the telephony revenue and bypass broadband service providers.
PacketCable telephony is supported by a number of vendors but is exclusive to the cable industry. While the PacketCable work proceeded and time passed, many companies proceeded in a different direction for packet telephony: using SIP as the call control protocol. Vonage uses SIP, and Microsoft provides native support in Windows XP and uses SIP in Windows Messenger. In discusions at Broadband Plus, several companies privately questioned PacketCable's market impact versus the many companies including Microsoft lining up behind SIP. The long time it has taken for the cable industry to develop and deploy PacketCable may provide an opening for Vonage and others to use SIP telephony to siphon significant portions of the packet telephone revenue away from operators.
The industry's perspective seems to be that Vonage and others might gain some limited base of those looking for cheap telephony and second line service, but that SIP telephony won't have a large impact without the high quality, reliability, 911 lifeline support and CALEA support in PacketCable. The industry also questions whether such providers would be able to scale to handle large numbers of customers.
QoS also provides an advantage for the MSOs. With services like Vonage, as long is there isn't congestion on the network, you can make perfectly satisfactory phone calls. But MSOs point out that problems can arise when the cable network traffic is heavy. What PacketCable can do that Vonage and others can't -- because SIP is not "network aware" -- is to use QoS to make sure that voice packets are given priority over other packets.
In addition, MSOs could deploy new traffic management tools, like those from Ellacoya, to provide priority to their premium service customers, such as those that buy the MSOs' cable telephony. Customers using SIP telephony over a standard cable data service would not be given priority and would receive poorer service, especially at peak calling times.
Craddock and Gough also pointed out that the business plans of providers like Vonage in the US rely on the so-called enhanced service provider (ESP) exemption for ISPs. This means that they are exempted from payment of certain interstate access charges. Changing this regulatory status might damage or destroy their business economics.
So what is the status of cable telephony? Many of those who have lived through all the industry and organizational changes since the AT&T/TCI acquisition are still at work to create the final deployed result. In 1998, Mark Dzuban was working on AT&T's cable telephony plans; today he is Executive Vice President, Cable Telephony Deployment, at Cedar Point.
In April 1998, Steve Craddock spoke on IP voice over cable technology in a session at the Voice on the Net (VON) conference. Since then he has overseen the development of the PacketCable specifications and organized trials in Union, N.J., Willow Grove, Pa. and Detroit. He and the Comcast team have already installed some of the equipment for residential VoIP phone service which will be introduced in a portion of the Philadelphia market during 2003.
We organized and moderated that VON session, now nearly five years ago. The long time serves as an object lesson in how long it takes from the time new technologies are conceived until they are operationally ready for deployment.
( www.attbroadband.com ) ( www.comcast.com ) ( www.ncta.com ) ( www.cox.com ) ( www.packetcable.com ) ( www.syndeocorp.com ) ( www.cablelabs.com ) ( www.arrisi.com ) ( www.cedarpointcom.com ) ( www.vonage.com ) ( www.net2phone.com ) ( www.libertycablevision.com ) ( www.microsoft.com ) ( www.ellacoya.com )