Before 1997, it was not possible to roam into or out of North America with mobile phones. Furthermore, it was also impossible to send or receive text messages across this continental boundary. The marketing vice president of Vodaphone, the largest mobile phone service provider in the world identified this problem, mentioning that hundreds of international flights arrived in the New York City area daily, and businessmen deplaning turned on their mobile phones, expecting service, but to no avail. Their only option for mobile service was to rent a mobile phone locally, but this was complicated. Not only was there a credit issue (prepaid was not widely available), but also a new phone number had to be used, and then communicated to the person’s contacts. However texting was still not possible. A similar situation existed with North American travellers overseas. This was not an acceptable situation , and Vodaphone wanted it fixed.
Several industry groups had been working on this problem, including a co-operative effort between France Telecom, Deutsche Telecom and GTE (US), but none were able to implement any practical and scalable solution. There were two main problems that had to be solved, a network specification incompatibility, and a network routing structure incompatility. A simplified explanation follows.
North American telephone signalling is governed by ANSI (American National Standards Institute), while most of the rest of the world follows the ITU (International Telecommunications Union) standards. While these standards have the same principles, their details diverge in important ways. The defined signals in SS7 signalling, used for telephony, have similar meanings between ANSI and ITU, but their bit pattern and fields are not the same. So any interworking interface must convert from one standard to the other. This is a fairly straightforward conversion that can be accomplished and developed using complex software. The more difficult problem is the routing structure. Both ANSI and ITU define various addressing schemes for message routing and node identification. Routing falls under classifications such as E.212, E.214 and E.164. As an example, everyone is familiar with E.164, as it is used to dial telephone calls, in the format Country Code, Area/City code and subscriber number. The way these fields are named and delineated can be on a national basis, but the structure remains the same. E.212 and E.214 have different structures.In North America, E.212 and E.164 routing is used, while most of the rest of the world uses E.164 and E.214, depending on message type.Because the assignment of subscriber’s numbers is on a national and telecom operator basis, there is no formula or rule that converts one format to another. Therefore, no software can accomplish this task. To overcome this obstacle, I developed the concept of a conversion table to perform the conversion. The challenge with this was not only technical, but logistical and commercial as well. The drawback of this approach is that every series of numbers specifying a mobile operator subscriber or node must be datafilled in the table, which has to be large enough to contain every range of every mobile operator and their subscribers in the world. Usually there are blocks of tens of thousands of numbers, each requiring a single entry specifying the beginning to end of the range. The show stopper is that there is no source or central database that lists all of these numbers (never mind real time updates). Each Mobile operator has his own, and would be required to populate and continuously update our conversion table, which in effect would become THE central database for the world. Enforcement of this requirement could be a problem, but it was a non issue, because if a Mobile operator wished to establish roaming with Canada or the US, he had to establish a commercial relationship with us, and provide his numbers, and updates. So the conversion table concept is very cumbersome, but the datafill task became automated with a sophisticated order entry system that was subsequently designed. to handle the huge table. The concept was presented to the GSM Association at a meeting in Zurich, Switzerland, and I was given the approval to implement the project for my employer, a Canadian network operator. Working closely with one of the most avanced telecom software developers, in Ottawa, the software and hardware were developed to my design specifications. The deployment of the International Roaming Signalling Converter (IRSC) was a huge success, both technically and financially, as all inbound and outbound mobile roaming and text messages into and out of the USA and Canada had to be routed through the duplicated converters in Montreal and Toronto, and message traffic eventually surpassed the billion messages per month rate.
The technical disadvantage of the table based approach actually became a marketing advantage, because even the smallest and most remote mobile operator had to become a customer if they wanted to roam or text with any American or Canadian operator, including AT&T, and T-Mobile. The customer base grew exponentially over the next few years, allowing the company to market many services.
The International Roaming Signalling Converter that I developed to perform the necessary interworking, was in service by 1997, and is still in service. It has been patented.in the US, Canada, and internationally.