Editorial : VoIP: Outlook for 2004 and Beyond

The general press, as well as most trade magazines, just can’t seem to get the picture right when it comes to the voice over IP (VoIP) opportunity. So here’s my take on the top 10 VoIP opportunities for the year 2004 and beyond. First, however, here’s a little historical perspective of VoIP and what fundamentally has occurred over the last couple of years that didn’t exist in the early years (1995-2000).

Most notably, today’s networks have lots of bandwidth to throw at the quality of service (QoS) problem that haunted VoIP in the past. In other words, there are VoIP opportunities when the networks those voice packets ride over are QoS-enabled. This can be accomplished with carrier-grade OSSs but hasn’t because service providers can just overprovision (e.g. throw bandwidth at the QoS problem, which is what is happening today).

Also, throwing bandwidth at the VoIP problem eliminates the need to use low bit rate digital coders (e.g. 8 Kbps coders vs. PSTN-grade 64 Kbps coders), and you don’t need to deploy silence suppression which causes clipped-off sentences, etc.

Finally, today’s public Internet backbone and/or managed IP networks offer excellent quality. However, three years ago a network service provider’s backbone typically lost 5 percent of packets and had a one-way latent delay of 100 milliseconds or more. Today, IP packet loss is down to 0.1 percent, and delays are approaching legacy TDM networks.

So what is the bottom line? Corporations now have 100 Mbps Gigabit Ethernet LANs, cable modems operate in the megabit per second range and the fiber glut has placed terabyte capacity pipes in the public Internet.

So whenever you have bandwidth to burn, you can find VoIP opportunities. Here’s my take on the top 10:

1) VoIP over DSL Here’s a classic winner and loser case. First the loser: the dozen or so VoIP service providers. These so-called start-ups that are selling flat-rate local and long distance voice services to DSL subscribers won’t survive in the North American market. Not to totally rehash the beating up we gave them in last August’s editorial, but in brief, they cannot control QoS on the outgoing link of the DSL service provided by the ILEC or CLEC. If the VoIP start-ups begin to get traction, their suppliers (ILECs and CLECs) will either shut them off because of service agreement violations or just reduce the DSL bandwidth.

The bigger problem the VoIP start-ups have is that they won’t have state PUC or FCC support in the long run. They can’t offer carrier-grade E-911, can’t support lawful intercept requests, pay nothing into the Universal Service Fund, pay no access charges, and more.

So where is VoIP over DSL a winner? It’s where you have two things: lots of bandwidth on the subscriber broadband connection and where the incumbent telco’s local rates are high and not flat rate. And where do you find such places? The Asia-Pacific region for starters. For example, Yahoo Japan has captured over 2 million NTT local customers with its VoIP over DSL service. Here are three reasons why:

1) Lots of bandwidth (10 Mbps to 27 Mbps)

2) High NTT rates: 10 cents for local vs. Yahoo at

3 cents, and calls to other Yahoo Japan customers are free.

3) Very favorable regulatory climate that pushes broadband residential access services at low DSL prices. Users pay $20 for DSL access including the ISP service.

Finally, what about “Skype”? This is a new company started by several programmers who wrote the software for Kazaa. If you can file share music peer-to peer, than why not VoIP? Skype allows free calls between two users who have downloaded their software. All the user needs is the software, a broadband connection like DSL and a basic headset, and you are ready to go. Of course, you have to understand peer-to-peer networking, and the called party needs a broadband connection as well.

So where’s the Skype-type VoIP opportunity for service providers and BSS/OSS vendors? Beats me. It’s free, and it will probably be ignored by the regulators. If the music industry failed to shut down Kazaa, then there is little chance for telco regulatory success. So as long as this VoIP service remains “geek-to-geek” it will not be regulated. But when you go “geek-to-grandma” you will have the same regulatory problems as the VoIP providers, which interconnect with the PSTN.

2) PacketCable At this point in time and well through 2004, cable companies will be sitting on their wallets regarding full funding of the PacketCable approach to VoIP over cable. Why? The softswitch approach doesn’t have the associated billing and OSSs that are scalable to millions of users. In addition, the cable companies don’t see the return of upgrading HFC networks to five 9s reliability and providing network powering of IP telephones. Finally, you have the alternative VoIP broadband model as described in opportunity 1 above.

The bottom line: The cable industry has the ingredients for VoIP over cable modems without massive investment to their infrastructure (softswitches, OSS, BSS, and so on). So here’s what they will do in 2004. Upgrade cable modem capacity to 3 Mbps (e.g. improve bandwidth and forget QoS OSS deployment), team with a local CLEC that has local circuit-switched infrastructure to support E-911, CALEA, access fees, etc., and offer second or teen line voice over cable modem. Unfortunately for BSS, OSS and network infrastructure vendors, it’s a cable company no-brainer for 2004 and maybe beyond.

3) IP-PBXs With 100 Mbps and Gigabit Ethernet plus MPLS-enabled IP backbone service offered today, you can match and exceed older TDM-based PBXs regarding service quality, reduced operational costs and new services. Avaya, Cisco and Nortel can’t produce IP-PBXs fast enough.

In mid-2003, IP-PBXs are selling at a rate exceeding TDM-based PBXs. It will take 10 years to see an 80 percent IP-PBX world, but it’s happening.

4) Managed IP Network Service The managed IP network service market consists of large multinational organizations, ISPs and wholesale IXCs. The providers like AT&T, MCI/UUNet, Sprint, etc., have enhanced their infrastructure with lots of fiber and MPLS routers that deliver high QoS for VoIP.

The most profitable customers are the multinational corporations. They need three network services: (1) Move IP packets carrying data from site to site (2) An IP transport to interconnect their new IP-PBXs and (3) Just plain voice transport between legacy TDM PBXs. The problem for the NSPs is that their traffic is doubling every year or so but their revenues from corporate customers just wanting IP transport are shrinking! Bottom line: VoIP by itself as a transport revenue generator is not the silver bullet that’s going to turn around the IXCs.

5) Hosted VoIP Services The positive side of this transition of enterprise voice to VoIP in the CPE and backbone transport space is that it creates opportunities for all facility-based carriers to move into the service management space. Big corporations want to outsource VoIP support for remote locations, and SMEs want to outsource all VoIP support. As new VoIP value-added services come out (instant messaging, presence, Web conferencing, etc.), they will create hosted service opportunities.

6) IP-Centrex IP-Centrex opportunities should be a no-brainer for ILECs since they have the majority of Centrex users, supported by their legacy TDM circuit-switched networks. If a customer leaves ILEC circuit-switched based Centrex for an IP-PBX or CPE solution, the ILEC loses big.

So what’s stopping the ILECs from aggressively pursuing IP Centrex? The answer for the most part is VoIP security. The vast majority of today’s large Centrex users are government organizations. They cannot afford to be taken down by hackers, if for fear of the negative PR alone. There are other issues like powering IP phones over LANs, etc. that have to be addressed before IP Centrex can take off.

7) VoIP over Wi-Fi Again here’s a case of winners and losers regarding VoIP opportunities. In the enterprise space, VoIP over Wi-Fi looks like a winner. If Wi-Fi makes sense as a wireless LAN as well as phones plugged into fixed line LANs connected to a IP-PBX, why not an IP Wi-Fi phone? Besides, the enterprise can engineer the Wi-Fi network for good QoS and security.

But what about VoIP at Wi-Fi hot spots? Other than softphones (software in a Wi-Fi enabled laptop with headset), VoIP phones or devices at Wi-Fi hot spots are a loser. Here’s why. First, you don’t have decent Wi-Fi coverage and probably never will outside coffee shops, hotels, airports, and so on. Remember the CT-2 venture 10 years ago and PHS in Japan? They failed because they couldn’t match 2G cellular coverage. But more importantly, you can’t guarantee or control QoS. Finally, ask anyone who has ever used a Wi-Fi hotspot if they had a cellular phone with them at the time. You will likely hear a near 100 percent yes.

8) SIP Phones Session Initiation Protocol (SIP) phones are the basis for today’s momentum of VoIP in the enterprise space vs. alternate protocols (H.323 and softswitches) available three years ago. The extrapolations out of the enterprise space to the public space are intriguing. The concept of plugging in a phone, getting a dial tone, performing self-provisioning of a new account, etc., is appealing. The opportunity is working out the ENUM roadblock (e.g. translating IP addresses to phone numbers and vice versa) in a global environment. Producing a $ 29.95 SIP phone will be orders of magnitude easier than solving the ENUM issue.

9) VoIP Over 3G True 3G mobile wireless has both voice and data going over the same channel as IP packets. With today’s 2.5G networks, voice signals travel over a separate channel producing TDM or circuit-switched type traffic.

There doesn’t appear to be any business case in North America at this time to implement 3G just to do VoIP. That being the case, the question of market demand may be answered with the rollout of CDMA-EV.DO (evolutionary, data only) by Verizon in the Washington D.C. and San Diego areas. For $80 a month, you get 400-500 Kbps data rates to your laptop. Also the EV.DO coverage is reported to be great in and around the D.C. beltway. If there is demand, and particularly for VoIP applications, that may be the pull true 3G needs. Otherwise why do VoIP over cellular?

10) Third-Party VoIP Applications If you want to get the attention of all telecommunications service providers, just announce as a third-party supplier that you have a killer VoIP application for their network and customers. No additional CapEx but a little OpEx is OK if there is a ROI in the current budget cycle. In short, is there an equivalent of caller ID on a circuit switch that generates $5 to $6 a month for an ILEC with virtually no additional investment other than software located in a service control point processor?

These are the best VoIP opportunities I see for year 2004 and beyond. Are there any big new OSS/BSS opportunities? Yes, but they are like gold in the ground: It’s deep, and you have got to be patient.

If you need to be brought up to speed on the latest technologies generating new business opportunities, plan to attend “Understanding Telecommunications for Non-Engineers.” For more information go to www.telestrategies.com.