The vision that the public switched telephone network (PSTN) will eventually be an all-VoIP network goes unchallenged in almost any newspaper or telecom trade magazine. FCC commissioners lead into the subject in public speeches by saying VoIP will increase competition, lower costs and introduce new services. Peddlers of VoIP market research reports point out that IP-PBXs are outselling legacy TDM PBXs, 20 percent of international traffic is VoIP and growing, and VoIP over broadband is growing as well.
Not to rain on the VoIP zealots’ parade again, but there are major challenges to overcome. Before I give you my latest top 10 challenges and how I see the road ahead to the all-VoIP PSTN, here’s why there are challenges.
Leadership
No one is in charge of the migration to an all-VoIP network. In the United States, ubiquitous circuit-switched PSTN is available to 99.99 percent of the population because AT&T or the old Bell System was in charge. You have the standard global Internet because the U.S. Department of Defense was in charge during its formative years. However, no one is in charge of the VoIP revolution.
Economics
The people who say VoIP will overtake all because it’s cheaper haven’t a clue about the total cost of telephone service. The reason VoIP has gained traction is because it reduces the cost of one element in a phone network, and that’s transmission. For long distance carriers, provisioning individual trunk or circuit groups, switch to switch, is inefficient and not necessary with IP transport. For corporate or enterprise network users, one transmission system (IP transport) can handle both voice and data more cost effectively than two separate transport systems.
But the argument goes that VoIP will generate altogether new services, which means new revenue. The reality is that there has yet to be one new service created by VoIP infrastructure that has generated any meaningful revenue or profit to date. When a corporation buys an IP-PBX, that’s new equipment revenue, not new telecommunications service revenue.
Stonewallers
The regulators have one thing right—VoIP will create more competition if only in the short term. As such, the transition to all VoIP is not in the best interest of Tier 1 service providers. RBOCs, in particular, would be better off if the FCC ruled “no local service on VoIP.” RBOCs are masters at stonewalling regulatory or legislative mandates if it means that they will lose customers to their competitors.
And it’s not just the RBOCs that have an interest in stonewalling VoIP. Government service folks also don’t take well to the unfunded mandates and other expenses VoIP will create.
So, keeping the issues of leadership, economics and stonewalling in mind, here are my top 10 challenges to an all-VoIP PSTN.
1. Access
Local VoIP only works if you have a broadband connection (DSL, cable modem, T1 circuit, etc). Yes, 25 million homes in the United States have a broadband connection, with a 20-percent growth rate. As for DSL, demand is driven by Internet access, not voice. Given the current market dynamics, my guess is you will see DSL access demand begin to flatten out at the 30 to 40 percent household penetration level. Why? A certain segment of the population can live with dial-up, they don’t have a computer in their house, they don’t have a phone line because they just use wireless, they live in a rural area, or they subscribe to cable modem access.
To get DSL into every home in the country, service providers would need an economic plan, a leader and no stonewallers. The Communications Act of 1934 came up with a plan to increase household phone penetration from 30 percent to nearly 100 percent today. AT&T (the old Bell System) was in charge, long-distance revenues were used to finance local service build-outs, and there were no stonewallers.
Regarding DSL broadband, there isn’t a leader or a plan, and there are stonewallers. The RBOCs provide DSL at a loss today, and every time a customer converts to DSL they often don’t see an increase in revenue because the customer disconnects their second phone line. Also, the RBOCs, with the exception of Qwest, won’t offer you DSL if you disconnect your circuit-switched voice service! But, cable modem access is different because cable companies don’t have to worry about cannibalizing another service. This among other things will catapult cable companies into the catbird seat over time.
2. Quality of Service
The quality of service (QoS) for VoIP is adequate because packetized voice only goes over one IP network. If VoIP in a public network sense takes off, the VoIP call will go from circuit-switched/TDM back to VoIP on a second or third network. Latency or delay, lost packets and voice codec incompatibilities will result in poor voice quality. The reality is that VoIP will work if the industry agrees to have only one VoIP carrier. Fat chance, except of course for the cable folks.
3. E-911
If there ever was a case of calling a pimple a beauty mark, it’s VoIP zealots presenting their case to the regulators and Public Safety Answering Point (PSAP) folks on why VoIP is an enhancement to E- 911 and worth embracing. People at the scene of an accident could send pictures to the PSAPs, contact E-911 with SMS text messaging and use sniffing devices to alert police to the type of chemicals spilled should a truck carrying hazardous materials be involved, etc.
The reality is that the E-911 system architecture is over 20 years old, is centered around ILEC-owned and operated databases (Master Street Address Guide [MSAG], etc.) and was paid for by the ILECs. It’s been a 10-year struggle to integrate location data from cellular 911 calls, and it’s still a work in progress. Dealing with VoIP is orders of magnitude more difficult.
Because you can move around with a VoIP phone (plug it into a LAN or Wi-Fi connection anywhere you want) and because assigned phone numbers can be foreign to the area you are dialing 911, it is difficult to determine the location of the caller. Currently, there are three solutions being pursued, but all three are problematic.
Solution 1: Forward the 911 call to the general number of the appropriate PSAP. This is how most VoIP over broadband providers are doing it today if they are not affiliated with a local CLEC. However, the 911 call typically goes to a PSAP receptionist and not directly to the trained PSAP dispatchers who have access to very sophisticated PSAP terminals and IT infrastructure. The PSAP community hates this solution because it introduces delay, and you don’t automatically get the location of the 911 caller.
Solution 2: Today, when you plug your VoIP phone into a LAN, you eventually get an IP address from a Dynamic Host Configuration Protocol (DHCP) server associated with that LAN or ISP if you are a roaming. Solution 2 would be to have that DHCP server also assign the location along with the temporary IP address.
However, DHCP wasn’t designed to notify location. Extending a protocol to do something it wasn’t designed to support is inviting a disaster. Second, I don’t know how many DHCP servers are out there. It could be hundreds of thousands or maybe a million. Finally, you would have to interface with the ILECs (think stonewallers) that control the MSAG servers. Solution 2 is great as a computer science Ph.D. thesis topic, but I don’t see it solving VoIP 911.
Solution 3: The third VoIP solution is even crazier. It calls for converting all PSAP infrastructures to VoIP. That’s 6,500 PSAPs in the United States! Bottom line, who would be in charge, where would the tens of billions of dollars come from, and what’s the incentive for the PSAPs and ILECs to go along with it?
4. CALEA
Currently, VoIP over broadband players claim they are “information service providers” and not subject to the provisions of the Communications Assistance for Law Enforcement Act (CALEA). The problem is that criminals will just choose telecommunications services, like today’s local VoIP, because of the weakness in wiretapping support. Bottom line, neither the regulators nor the service providers are stepping forward to make VoIP networks fully CALEA compliant. The technology is there in part, and the gaps could be filled given equipment demand. In the long run, law enforcement will win on this issue, and VoIP providers will have to fall in line.
5. Security
For all practical purposes, denial-of-service (DoS) attacks on the PSTN can’t be launched from the outside. You can say the same thing about today’s VoIP backbone transport networks because they too are closed except for phone access. But when it comes to an always-on VoIP phone, look out for the obvious things to expect from hackers: DoS attacks, hijacking your phone calls, your number, and so on. Note, the only reason you don’t hear about security breaches in VoIP over broadband today is because information service providers don’t have to report these problems to the FCC. Bottom line, all the security problems that occur on IP networks today will occur on VoIP networks and more.
If hackers can take down the ultra technically sophisticated Akamai, which manages 15 percent of the world’s Internet traffic, for a couple of hours as they did last month, then think what a hacker could do to an all-VoIP PSTN.
6. IPv6
If you are going to see an all-VoIP network with convergent services, every phone will need a static or permanent IP address just like the phone in your home has an assigned telephone number. The problem is there are not enough IP addresses under the current IPv4 scheme. Note: IP addresses are 32 bits, which produces in theory 4 billion numbers. They are allocated in blocks, which results in assigned but unused numbers. So even today, the 1.2 billion wireless users in the world couldn’t be assigned a static IP address.
Today, there are gateway solutions between legacy IPv4 address devices and the newly emerging IPv6 address devices. However, no one knows how to manage the migration to all IPv6, let alone who would be in charge of the process. It will be expensive, and today’s U.S. carriers that hold the lion’s share of numbers don’t have the financial incentive (think stonewallers) to migrate to IPv6.
7. ENUM
On the road to an all-public VoIP network, there will be both IP addresses and plain old telephone service (POTS) numbers. A directory server is needed to translate numbers from IP to phone numbers and vice versa, just like DNS servers translate domain names into 32-bit IP addresses of 1’s and 0’s today. That’s where electronic numbering (ENUM) protocols come in and the look of a server takes shape. In order to have a true public VoIP network with many service providers all interconnected to the PSTN, you need one ENUM directory.
Here is the challenge. On the one hand, being in charge of a national ENUM directory puts that entity in the catbird seat regarding new VoIP business opportunities. Given there are nearly 3,000 service providers in the United States, how do you get these guys to cooperate? On the other hand, you have the SIP phone zealots with SIP proxy servers that point out that an ENUM directory is not needed in an all-SIP environment.
The bottom line is with private ENUM and/or SIP proxy servers, you can roll out VoIP. It will become problematic if you get to several million users scattered in multiple VoIP network islands without bridges to other islands. It would be like: “I would like to call you, but you can’t be found in my service provider’s ENUM directory.”
8. Standards
If there ever was truth to the statement, standards are great because there are so many to choose from, it is here. Not only are there hundreds of standards groups addressing the many facets of VoIP, they tend to address only the politically correct issues. Things like security and CALEA are virtually ignored. And it’s worse than that. If you really don’t want standards to evolve (think CALEA), you go standards body shopping. Just like class action lawyers know which state courts will be sympathetic and choose that state to file their case, service provider associations do the same.
9. Training
How many telco employees understand VoIP, or for that matter IP protocols and technology? Consider the recent labor problems at SBC. They outsourced customer service for their DSL customers overseas. My understanding is their argument was that the skill set wasn’t available in-house. Now they have a contract that says they will provide work for life for 100,000 craft labor union workers. Migration to all VoIP will require considerable training and financial resources. How do you make the ROI case on this investment?
10. Taxes
Again relying on an old saying, “In this world nothing is certain but death and taxes.” Taxes are a big thing in telecommunications. On average, telecom taxes are 18 percent of your phone bill. Only cigarettes, alcohol and gas have higher taxes.
Here are the dumb things that VoIP people say about taxes. Our services are lower in cost because we don’t charge taxes. We don’t need granular billing systems that gather call records because our services are flat rate and distance insensitive. Or we will just sort of guess what the taxes should be and err on the high side and just pass the error off onto the consumer.
Here’s the reality. Almost every billing system will have to be upgraded to handle VoIP taxation. You can’t charge too little because the tax guys will get you at audit time. So you have to produce granular CDRs reflecting intra-state, interstate, call duration and more. You can’t err on the high side because you will become subject to class action lawsuits. Also, you can’t say a packet is a packet regardless of whether it is voice or data because to the tax people, voice and data are taxed at two different rates.
Finally, the tax people don’t care what the FCC does. If it looks like a voice call, they will tax it like a voice call. Only Congress and state and local officials have a say on what can be taxed and what can’t.
So Where Is VoIP Headed?
The good news for VoIP, particularly VoIP over broadband is that it is the only competitive alternative to local service. By not supporting UNE-P and appealing the U.S. District Court’s ruling on TELRIC to the Supreme Court, the Bush administration has taken the teeth out of the local competition portion of the Telecom Act of 1996. If the FCC wants to support local phone competition, it has to be pro-VoIP.
The bad news for the all-VoIP PSTN vision is that local or Class 5 switches won’t disappear for at least another decade. Yes, you will have VoIP access on the customer side and VoIP on the backbone, but nothing is quite ready to replace the switches that sit in between. Why? E-911, CALEA, billing, taxation and other BSS/OSS infrastructure is still a work in progress.
If you are a part of this work in progress and must address the challenges listed above but need more information, plan to attend our new seminar, “Understanding Carrier Grade VoIP Technology and Back Office Requirements.” Or if you want to meet the BSS/OSS/ISS vendors that have the VoIP solutions, plan to attend TeleStrategies’ VoIP Conference and Exhibition on Nov. 8-10, 2004, in McLean, Va. For more information go to www.telestrategies.com.
Email this article
Add a comment
Printer version
Order reprints
RSS Feed
Bookmark article