Billing World Publisher's Letter On the Road to SS8

With every new generation of voice switches comes a new generation of switch-to-switch signaling protocols. In the 1950s electromechanical crossbar switching came along and so did MF or CCITT#5 signaling. The 1970s saw digital toll switches (4ESS) and also CCIS or CCITT#6, and in the 1980s digital local switches arrived and so did SS7 or CCITT#7. As we approach the year 2000, a new voice switch is on the horizon-the IP Gateway switch. Therefore a new signaling system must be just over the horizon as well-called SS8-right?

Well yes and no. The IP telephony camp is standardizing the digital signal processing functions in IP Gateway switches such as analog to digital coders, echo control etc. But other than facilitating interfaces with SS7, the IP telephony crowd is not openly working on a replacement for SS7. On the other hand, neither the RBOCs or telecom engineers are seriously addressing the next generation switch to switch protocol. Why not? Consider the following:
1. SS7 infrastructure in place represents tens of billions of dollars in investment.
2. SS7 is nearly a true global standard.
3. SS7 is rich with new applications and opportunities (see Paul Gowans's article on page 28)
4. SS7 is the foundation technology of a major regulatory mandate called local number portability.
5. SS7 makes the carrier's network intelligent, profitable and indispensable.

The bottom line is the powers that be (RBOCs, large IXCs, foreign PTT's and regulators) have no incentive to make SS7 obsolete. But the next generation of telephony-IP telephony over the Internet, or more important-voice over intranets-is coming like a freight train that can't be stopped, and SS7 is nowhere near an ideal signaling system to interconnect IP Gateway switches.

Five Forces Driving SS8.

Here are five forces I see driving the inevitable creation of SS8:

1. IP Gateway Switches.
If it were up to the world's established carriers and voice switch vendors, IP Gateway switches would never see the light of day. But the new generation carriers such as Qwest, with its super-high-capacity national fiber networks, 7.5 cents per minute voice over IP (VoIP) calling rate and a newly acquired LCI's voice customer base, are creating a new kind of competition. Behind Qwest are Level 3, IXC Communications, Williams and hundreds of international Qwest wannabes with lots of fiber capacity but no customer base to cannibalize with IP telephony. Also in the wolf pack are router vendors (Cisco, Bay Area Networks, and so on), that don't have voice product lines to be made obsolete with IP gateway switches, and whose CEOs stare at staggering growth opportunities with circuit switch replacement.

2. New IP Telephony.
The only reason IP gateway switches need SS7 is because today's telephones signal with touchtone or dual tone multi-frequency (DTMF) signaling. If a VoIP carrier wants to avoid giving their customers a second dialtone and asking the customer to enter a PIN number for billing plus the called party number, the IP gateway switch must interface with the local exchange carrier via SS7 similar to the way long distance switches do.Just as a new generation switch makes the last generation switch-to-switch or trunk signaling protocol obsolete it also makes the last generation telephone access or line signaling protocol obsolete. Crossbar switches, introduced in the 1950s, were accessed by rotary dial or dial pulse (DP) signaling. Analog computerized switches (1ESS) introduced touch-tone or DTMF signaling in the 1960s, and digital switches of the 1980s introduced ISDN. With IP Gateway switches will come a new generation of line signaling, and it will be IP based. I guarantee it or your subscription to Billing World will be free.

3. New Access Technology.
Every CLEC's marketing dream is to put a box in a office building linked to their SONET or SDH fiber ring via a fiber optics pipe carrying bits at OC-3 (155 MBPS) or OC-12 (622 MBPS) rates or higher. This box will interface with building tenants' voice, data, or video terminals and it will likely be an IP gateway switch. SS7 will not service this OC-3/12 access pipe and IP gateway switch very well.

4. Running Out of Numbers.
The original North American Numbering Plan (NANP) uses NPA-NXX-XXXX-a ten-digit number-to identify a telephone, the first six digits (NPA-NXX) to identify a switch and the first three digits (NPA), to identify an area. The plan, introduced in 1947, reflected the voice switch capabilities of the day (Step-by-Step switches) with a 10,000 line capacity and the requirement that a Numbering Plan Area (NPA) have a 0 or 1 as the middle digit, thus limiting area codes to 160 possibilities (actually 152 subtracting 911, 411 and so on). This plan was supposed to last into the 23rd century, but it did not. Note that in 1947 there were no cellular phones, pagers, fax terminals or second phone lines for Internet access. The NANP was changed in 1995 to create more numbers. By then the old Step-by-Step switches were long gone, replaced by computerized switches that can handle area codes with middle digits 2 through 9. This increases area codes by 640 and telephone numbers by eight million per new area code. Yet, at current rates, these new numbers will be used up early in the 21st century or even within five years.

Today two out of three new telephone numbers in the United States are used for wireless voice services (see Billing World, April 1998). Just wait until the new wireless data devices arrive. In the near term they will be accessed by dialing a ten-digit NANP number, not an IP address. If that doesn't exhaust numbers, geographic number portability will. What is going to stop a florist who sees a competitor flash 800-FLOWERS on the TV from getting a local FLOWERS number in every NPA and treating it like an 800 number? (Run through all the vanity numbers BAGGAGE, HOLIDAY, WINDOWS etc. and you will see the number stockpile quickly dwindle.) Finally, blue-sky things like talking and listening appliances that can receive calls will appear and more. There are endless ways to attack the number pool. Expanding or manipulating the ten-digit number plan will make SS7 obsolete as well as the hundred of billions of dollars of software based on ten digit telephone number fields.

5. Running Out of NPA-NXXs.
The SS7 numbering system is based on the area and office code, or NPA-NXX, and before telephone numbers run out, we will first run out of NPA-NXXs. Why?
The days when the NPA-NXXs were assigned to the RBOCs, who in turn assigned blocks of 10,000 numbers to other carriers, are over. Today anyone can become a CLEC and soon these CLECs too will be assigned NPA-NXXs. In addition, just as the practice of one company-one telephone number was made obsolete by automatic PBXs and Direct Inward Dialing, the concept of one household-one telephone number and no premises switches also will become obsolete.
The SS7 architecture was not designed for multiple carriers gobbling up blocks of numbers to be assigned at a later time to their users. Eventually IP telephones and devices will have IP addresses. But for now the new CLEC VoIP carriers will have their own NPA-NXXs.
Does local number portability solve this problem? A carrier or pseudocarrier could get numbers out of another carrier's pool that could be assigned to the initial user, right? Not really! The problem with this is economics. Forget about the fact you have to pay the real holder of the number for the database dip to get the location routing number that will direct callers to your switch. The real problem is service activation: It's expensive (millions of dollars) to link into a Number Portability Administration Center, and it's not Point-of-Sale friendly. Real-time access or services activation from someone else's block will not be available in the immediate future. Carriers will want their own NPA-NXX blocks just like the RBOCs, and there aren't enough to go around.

What Is SS8 Going to Look Like?

Challenged by these problems, SS7 will become outmoded. Here's my vision of the next generation of inter-switched protocol or SS8.

1. IPv6-Based.
It's a no-brainer that SS8 will ride on TCP/IP. Why? Next-generation telephony is being driven by the Internet, which is TCP/IP based; TCP/IP is cheap technology; a lot of applications development tools and programmers are available; and it's an even more globally accepted standard than SS7. It's the only modem computer internet working standard with global acceptance. However, today's TCP/IP is based on IP version 4 (IPv4), and it too will be replaced by IPv6. (IPv5 didn't make it and no one I know remembers why.) There are problems with IPv4: We are running out of IP addresses. IPv4 has 32-bit address fields broken up into four eight-bit blocks. The first three blocks are called A, B and C class license respectively. Although 32 bits yields 4 billion addresses, users get a block of IP addresses when they get an A, B or C class license. For example a C class license allows up to 256 addresses for internal assignment within your network even if you don't use them. So just as with telephone numbers, class licenses run out before IP numbers.

Over the last several years the Internet community has started packing IP numbers into smaller blocks, half or an eighth of a C class block for example; thereby IP number exhaustion has been pushed out for a couple of years.

IPv6 has 128-bit addresses. Enough IP addresses for every electron in the universe, I suspect. But half of the address fields will be used for identifying the traffic flow (data vs. voice vs. video); identifying other enhanced features, including the identification of latency delay sensitive traffic such as voice; and addressing minor items (joking, of course) such as billing. It seems logical to me, and therefore it must be logical to the rest of the telecommunication, computer and content provider industry, that if both SS7 and IPv4 need to be replaced, then kill two birds with one stone. Utilize IPv6 addresses for telephone numbers and Internet addresses!

2. No Separate Line Trunk Signaling
Today touchtone or DTMF signaling is used on the line side of a voice switch and SS7 on the trunk side. Why? When DTMF was invented (1960), digital circuitry was expensive, thus phones had limited functionality. Also fraud-control tools were minimal, so you had to protect the billing activation signaling system. Looking ahead, not only will the next generation phones transmit voice and data in IP packets, but also include the features of today's voice-based intelligent network infrastructure (packet switches or STPs, databases or SCPs, and service creation interfaces or SCEs). The next generation large-user networks will look more and more like today's carrier networks. Bottom line, SS8 will handle access signaling to and between switches, furthermore no separate access signaling protocol will be needed.

3. Microsoft NT or Sun Java Based
Either Microsoft or Sun (or both) and their respective partners will probably determine what the next generation phone is going to look like in the brave new world of SS8. Microsoft is making a big push in telecommunications (see Ed Finegold p. 18) and obviously would like to dominate the communications platform market. (Consider this: when was the last time Bill Gates started something new on a small scale, or for that matter failed at something?) Microsoft has a vision of a programmable network and probably a vision about what SS8 should look like as well. On the other hand, Sun has not been ignoring next-generation telecom networks with its focus on Java network computing and legacy telephone system access (see Billing World, October 1997). Lots of today's billing and customer care vendors support Sun in its goal to get Java to the desktop.

Before specifying and standardizing SS8, there must be at least a partial industry consensus, on what the standards should be, and the consensus has to come from the computer side as well as the telecommunications side of the table.

Finally, where will SS8 first appear? Will it be proprietary? Will it come from a forum or as a government-sanctioned standard? Here's how I see SS8 surfacing. Whatever telecommunications carrier lands a set of big multinational customers with its VoIP, Intranet offering and a desktop network computer solution will create a network model for other IP carriers. Alternatively, an IP carrier who lands a major industry consortium with an extranet offering-such as for the auto industry with all their suppliers on one IP network-will also be in a good position to define SS8 for the world. Regardless, SS8 will be created to support private networks carrying voice, data and video over IP technology, then public network SS8 will follow.

The good news for the Billing World reader is that no matter what SS8 looks like, next-generation networks will need new or upgraded billing, customer care and OSSs. We will keep you posted on the road to SS8.