Hairpin Curve

Calls to roamers usually end up routed in a circuitous, inefficient hairpin. So why don't calls just take the shortest path from point A to point B?

Just about everyone with a wireless phone can go roaming through the gloaming — and in the morning, afternoon or night, as well — without being aware of how complex the process is. Landline phone systems don't have to worry about their phones' whereabouts. Wireless phones, however, continually have to report their whereabouts because this information is needed to route incoming calls to them.

Mobility management is an essential part of roaming, at least if roamers want to be able to receive calls and not just originate them. This process is simplified by keeping information about the mobile's location in only one place: the HLR. Because any telephone system can find a mobile by querying the HLR, locating a mobile is a 2-step process: ask the HLR where the mobile is, and then route a call there. Because an HLR can be located based on the MIN, international mobile subscriber identity (IMSI) or directory number associated with a wireless phone, there's no difficulty in doing this.

Unfortunately, this process has a glaring inefficiency for roamers because calls also are routed through the home system. When a wireless subscriber is in a home system, this approach has little impact, but when he is roaming elsewhere, it can increase significantly the trunking required to set up a call. For example, if an American subscriber is in Argentina, and a Argentinean colleague calls him, the call will be routed from Argentina to the home system in the United States and then back again. A call across the street has been turned into two international, long-distance calls.

This inefficiency also occurs within a country. Calling a user's wireless phone from a pay phone in Miami results in a call from Miami to Calgary and then to wherever he's currently located — which could be across the street in Miami.

Is it possible to eliminate this inefficiency? Yes, but only a qualified yes. Even if providers started working on solutions today, it would be several years before implementations were widespread. Why bother to eliminate this inefficiency if it's so difficult? After all, roamers still receive calls, and the complex routing is transparent, so what difference does it make?

There are several reasons why consumers would benefit from optimizing routing to roamers:

• Reducing trunk usage would reduce the cost of calls to roamers. This cost is either directly charged to the subscriber or built into the cost of a single-rate plan.

• Calls may take considerably more time to set up with two unnecessary long-distance legs.

• The more trunks involved in the call, the more likely it will be refused due to a lack of trunks.

• Call quality may suffer, particularly if not all facilities are digital.

From the provider's viewpoint, optimal routing could make it more competitive through reduced prices and could increase the quality of service that it could offer to its subscribers when roaming.

If optimized routing to roamers is so worthwhile, why is it so difficult? There are several reasons: numbering difficulties, wireless/landline integration issues and international protocol-interworking issues.

Numbering Difficulties
If our caller in Argentina is using a pay phone, how would the local switch know that he's calling a mobile phone? All the switch sees is an international dialing access code, a country code and a bunch of digits that it doesn't interpret. Even the toll switches in Argentina are unlikely to look any further through the dialed digits than the country code. The Argentinean switches would need a complete database of U.S. area codes and office codes to make this distinction and, with LNP, even this wouldn't work.

But there is a potential solution. The SS7 ISUP standard has defined a new capability known as release to pivot (RTP), which allows a call to be set up to one destination and then yanked back and set up to another destination. This feature is designed for services such as enhanced 411, where the call can be set up to the number provided from the directory — at an extra charge, of course. Forwarding the call from the directory-assistance center is possible but would tie up all of its trunks, whereas RTP allows the directory-assistance trunk to be re-used once the call is diverted.

In the optimal-routing context, the Argentinean pay phone would initiate a call to the wireless phone's home system, which would divert the call to a phone number, known as a temporary local directory number, provided by the visited system in Argentina. Then the home system could drop out of the call, which would be re-routed by the local switch connected to the pay phone to the appropriate Argentinean wireless switch.

Wireless/Landline Integration
Unfortunately, there's still a problem that stems from the lack of coordination between wireless- and landline-standards bodies. RTP was designed to support landline services, without consultation with wireless-standards bodies. As a result, it comes close to meeting the needs of wireless — but not close enough.

The difficulty is that there's a period during wireless call setup when it's not known where the call will end up. If the roamer in Argentina didn't answer his phone or if it was busy, the Argentinean landline switch wouldn't know how to forward the call appropriately.

What's needed is the ability to maintain two call legs until either the roamer answers his phone or a call-forwarding condition occurs. If the roamer answers, the leg to the home system can be dropped safely. If a call-forwarding condition is detected, the leg to the visited system could be dropped, a new leg could be established to the call-forward number, and, if further call forwarding isn't possible, the leg to the home system could be dropped.

Protocol Interworking
Telecom standards suffer not only from parochial divisions into landline versus wireless — with the critically important SS7 standard being artificially pigeonholed as "landline" — but also from being generally defined on a national basis. So even though ATIS T1S1, a U.S. standards body, has defined RTP for ANSI SS7 ISUP, other national SS7 protocols may not yet support it or may support some variants.

If telecom providers make the jump from SS7 to IP as the underlying transport protocol, they have an opportunity to gain one method of interconnecting telecommunications systems that works between countries and that works equally well between wireless and landline systems. This could enhance the level of services that consumers receive and increase the acceptability of wireless services. Trying to standardize and consistently implement a capability such as RTP in multiple national SS7 variants would be much more difficult.

One of the biggest overlooked issues in telecom is the lack of wireless/landline integration. This gulf is the reason why calling party pays has never been widely deployed in the United States and why operator services are generally not available from wireless phones. It's also why, for a phone call to get across a camino in Buenos Aires, it may have to travel to the United States and back again.

Crowe (crowed@cnp-wireless.com) is a wireless-standards consultant and editor of Cellular Networking Perspectives, a wireless-standards and -technology bulletin.