[ Home | Glossary | Acronyms | Links | Contact us ]

Cellular Networking Perspectives

David Crowe’s Wireless Telecom Magazine Articles

Q1’2001 Issue

International Roaming

Many Canadians who have tried to use their phones while roaming outside North America may agree with Lord Alfred Tennyson, “Our home is far beyond the [radio] wave; we will no longer roam.” But, they might be wrong; international roaming is possible more often than most people realize, and is surprisingly economical ... when it works. Unfortunately, just as often, customers afflicted by Murphy’s law find that they just happen to have the wrong type of phone technology, or the right technology but the wrong frequency, for wherever they are roaming.

Seamless international roaming is a very complex problem to solve, requiring appropriate standards, interworking technologies and resolution of numerous business problems. However, one by one, the barriers are being removed, and two by two, Canadians are learning that taking their wireless phone outside their country is a smart thing to do. In fact, forward-thinking carriers realize that international roaming has a huge, largely untapped potential for new revenue, both from calls made by their Canadian subscribers roaming to other countries, and from servicing foreign customers roaming into Canada.

Taking the MIN to the Max

GSM technology arose in an international environment, so it has fewer technical barriers than North American radio technologies (TIA/EIA-553 analog, ANSI TIA/EIA-136 TDMA and ANSI TIA/EIA-95 and IS-2000 CDMA) that grew up in the US and Canada (which, for telephony purposes, are a single country) and that use the ANSI-41 network for a backbone. Whereas GSM phones are identified by the International Mobile Subscriber Identity (IMSI), the TIA family of standards uses the Mobile Identification Number, based on North American national phone numbers.

The MIN is 10 digits long and is usually programmed in North America, with a national phone number (i.e. without the “1” country code). Other countries followed suit. Mexico, for example, initially programmed their phone MINs with numbers of the format “52” plus their 8 digit national phone number. Imagine their shock when the area code “520” was assigned to Arizona. Realizing that their ability to roam into the United States was threatened by this ambiguity, the Mexicans pressed for a solution, choosing to use MIN codes beginning with the digit “0” or “1”, which are unused by North American carriers.

There still was a need to ensure coordination between TIA technology users outside North America, and eventually the International Forum on ANSI-41 Standards Technology (IFAST; www.ifast.org) was chartered to assign unique identifiers to international carriers. This need was highlighted when it was discovered that several North American wireless data systems (e.g. Aeris, Cellemetry and UPS) were also using this resource in an uncoordinated fashion.

The use of an International Roaming MIN (IRM) ensures the global uniqueness of every mobile identity, allowing TIA compatible systems around the world to recognize where roamers are coming from (i.e. the address of the HLR).

IMSI - Identifier of the Future

The International Mobile Subscription Identity (IMSI) has always been used by GSM systems, but over the past few years has also been worked into ANSI digital cellular/PCS standards (TIA/EIA-136 TDMA and TIA/EIA-95 CDMA) as well as the TIA/EIA-41 network backbone. This 15-digit identifier not only allows the identification of more phones than the MIN, but it explicitly identifies their home country in the first 3 digits (MCC, 302 for Canada). Consequently, each country can manage their own IMSI numbering plan without fear of conflict.

GSM systems do have a problem with the use of IMSI in North America. The standard has only recently supported 3 digit Mobile Network Codes (MNC) that are required in North America due to the large number of independent licenses. Not assigning a separate MNC to each license will necessitate reprogramming or replacing SIM cards if a license is sold or traded.

Implementation of IMSI in TDMA and CDMA systems has been slow, largely because of the installed base of phones that do not support it, some models of phones that do not support IMSI properly, and lack of any support when phones are in analog mode. Implementation of IMSI is a case of short term pain for long term gain, and so far TDMA and CDMA carriers have exhibited a low pain threshold.

Figure 1: International Mobile Identifiers

Terrible TLDN Troubles

The TLDN (Temporary Local Directory Number, known as the Routing Number in GSM) is critically important to call delivery to roamers. This number is assigned from a pool by the system servicing a roamer, and allows delivery of a call from the roamer’s home system through the landline telephone network. Most North American systems only use nationally formatted TLDNs which makes it difficult to route calls internationally. Complex tables will have to be maintained until the TIA/EIA-41 carriers can all migrate to internationally formatted TLDNs.

SS7 Signaling

The SS7 network is the nervous system of wireless carriers around the world. SS7 has a major limitation, however – it is a collection of similar, but not identical, national networks. The major address type in SS7, a point code, is only national in scope and cannot be used to transmit GSM MAP or TIA/EIA-41 messages across international boundaries.

One solution to this is essentially a kludge – national point codes are assigned to carriers on another national network. A Brazilian carrier might be assigned a Canadian point code. Sending an SS7 message to that point code would actually result in transmission to a specialized gateway that could map the Canadian point code to a Brazilian point code and reformat the SS7 signaling message to Brazilian standards. Another solution is to extend ANSI SS7 network links into other countries. For example, at one time, Hong Kong wireless systems were, according to the SS7 network, located in Kansas! This use of point codes is inefficient, and results in maintenance of additional mapping tables.

The long term solution to SS7 routing is to use a secondary address known as a global title. Any type of telecom number can be defined as a global title, and then routing will be based on it rather than the point code. In a wireless context, Mobile Directory Numbers can be used as global titles to route calls or short messages to a mobile, and MIN or IMSI can be used to route messages from a Serving MSC to an HLR (Home Location Register) or Message Centre (MC).

If SS7 addressing is compared to the IP addressing that many people are familiar with, the point code is equivalent to an IP address, and the global title is equivalent to a domain name. Beyond that, however, the analogy breaks down. IP is truly a global network, although it has many deficiencies when it comes to routing telephony traffic, particularly in its lack of robustness, security and guaranteed performance. Because IP is a global network, it can translate a domain name directly into an IP address, whereas this is not directly possible with a global title. A network point sending an SS7 message via a global title would not be able to use the destination point code if it was in another country, for example.

Global title routing has been implemented to some extent in GSM, but hardly at all in TIA/EIA-41 networks. GSM has attempted to simplify the problem by using phone numbers according to ITU-T recommendation E.164 as their only global title. This requires a mapping from E.212 IMSI to E.164 that breaks down in the North American Numbering Plan where numerous MCCs (e.g. 302 for Canada, 310-316 for the US, and 330-376 for the Caribbean) map onto a single E.164 country code (“1”). The ITU-T specified mapping between these country codes (E.214) assumes a 1:1 correspondence between MCC and CC.

The widespread implementation of multiple global titles is a major technical and coordination problem. Every STP (SS7 Signaling Transfer Point) has to be programmed with routing tables for each different global title, and every national SS7 standard has to have a compatible set of global titles defined. This process is complicated by local number portability that has almost doubled the number of global titles that are required. Furthermore, global titles are required for different applications (such as routing to an HLR and to an MC). Multiplying the number of distinct global title types by the number of STPs and international gateways that will be required in the global wireless signaling network reveals an extremely large distributed database maintenance problem!

If global title routing is implemented internationally, wireless entities in one country (e.g. an MSC in Canada) will route messages to an international gateway using a national point code. This device will interpret the global title, reformat the SS7 layers, and determine the point code that is the destination address of the wireless entity in the other country (e.g. an HLR in Mexico).

Figure 2: International SS7 Signaling for International Roaming


So far, we have just discussed how to make international calls work. There is no point in doing this if nobody is going to get paid for the calls. GSM systems usually produce billing records in the TAP format and TIA/EIA-41 systems in CIBER. As long as roaming is not between technologies, and only between countries, differences between these formats are not important. Both formats handle multiple currencies, one of the biggest issues with international billing.

Billing records are generated by the billing system of the MSC serving an international roamer, and then transmitted to the billing system of the home network, that will actually bill the mobile subscriber. A process called net settlement adds up how much each carrier serving roamers owes each other carrier that will bill the roamers. Each carrier then only pays or receives a single amount from the bank performing the net settlement, removing a major administrative nightmare when dealing with roamers from many different countries.

Future Directions

One of the inefficiencies of both GSM and TIA/EIA-41 becomes extreme with international roaming – inefficient routing. If a Canadian subscriber roams to Japan, and a Japanese friend or colleague phones them, the call is routed to Canada and back. Two international long distance calls to phone across the street! Solving this problem is not as easy as identifying it, requiring a level of integration between landline and wireless systems that does not exist today.

Another direction of telecom is inexorably towards an IP backbone. Internet protocols resolve the international roaming problems that are related to SS7, but unfortunately the protocols are not yet in place to allow carriers to feel comfortable leaving SS7 behind. A considerable amount of standardization work lies ahead before this goal can be reached.

Should I take my Phone to Elbonia?

With all the problems that exist, it might seem that my recommendation is to leave your phone at home when you travel. In reality, many of the international roaming problems have been solved – by interim solutions, perhaps, but solved nonetheless. TDMA, CDMA and analog phones should have no problem working throughout the Americas. If roaming does not work for you, phone your carrier and tell them. National roaming only became a necessity, not a luxury, when customers started to demand it. The same will happen with international roaming.

GSM phones in North America use different frequencies than in other countries, so only SIM roaming is possible. A GSM phone can be rented in another country, and the SIM will ensure that billing is performed by your home carrier.

In cases where international roaming does not work, you can use one of several services (such as WorldCell) that will rent a phone guaranteed to work in the countries you are travelling to. And, if that still is not good enough, or you are travelling outside cellular coverage areas, there are always satellite phones (e.g. Globalstar) that will work just about anywhere in the world.


Your name:
Your email address:

© – Copyright Mon, May 14, 2007: Cellular Networking Perspectives Ltd.