Dial-Up Roaming: 1995-2010

For roughly a 15-year window from the mid-1990s through the late 2000s, dial-up Internet roaming was the dominant connectivity product for international business and leisure travelers carrying laptops abroad. A roaming-dial-up account let the customer dial a local-rate access number in any participating country, authenticate as the home-account subscriber through a credential-proxy chain, and receive an Internet session billed to the home account — rather than incurring punishing international-long-distance charges for dialing back to the home ISP's POP, which was the only alternative.

Tempest Telecommunications operated one of the largest commercial dial-up roaming networks of the period, covering 150+ countries through a combination of direct points-of-presence (POPs) and partner agreements. This page documents the broader market history of dial-up roaming and Tempest's specific operations within it.

Commercial dial-up Internet emerged as a category in the early 1990s following the 1991 commercialization of the US Internet backbone and the parallel opening-to-commerce decisions in other major markets (the UK in 1992, Germany and France in 1993-1995, Japan in 1992-1993). The basic architecture was simple: a customer's PC ran a modem that placed a local telephone call to an ISP's modem bank ("points of presence" — POPs); the modem call carried PPP-encapsulated IP traffic; the customer's session was billed per minute (or, after 1996 onward in major markets, per month at a flat rate).

The dial-up era's peak years were roughly 1998-2003 in the major Western markets, with global dial-up subscriber counts measured in the hundreds of millions by 2000. The US dial-up subscriber base peaked at approximately 35 million households around 2000 (AOL alone counted over 30 million subscribers); the major European markets followed similar trajectories.

The dial-up decline began in earnest from 2001-2003 onward as ADSL broadband rollout reached residential markets across Western Europe, North America, and the more-developed Asian economies. By 2007-2010 dial-up was a residual product in most OECD markets, persisting longest in rural or geographically remote subscriber bases and in markets with slower ADSL rollout. The roaming-dial-up category collapsed across approximately 2007-2012 as universal mobile data made it obsolete.

A traveler in 1998 carrying a laptop wanted Internet access in their hotel room or airport lounge. The options were limited:

• International long-distance to home ISP: Dial a number back home (your own ISP's POP) over the hotel's PSTN line at international long-distance rates. Could cost $3-$10 per minute at era-typical hotel-PBX markup rates. Operationally painful.
• Buy local Internet access: Sign up with a local ISP in the country being visited, paying their setup fees, getting a local account, and using it for the duration of the trip. Operationally painful and impractical for multi-country itineraries.
• Hotel business-center kiosks: Available at major hotels, useful for emergency email, expensive per-minute, no offline workflow.
• Roaming-dial-up account: A specialized product where the customer's home ISP partnered with regional ISPs in destination countries. The customer dialed a local-rate number in the destination country and was billed by the home ISP for the Internet usage at agreed per-minute rates, with the home ISP settling separately with the destination-country partner.

Roaming-dial-up was the product that solved this problem at the technical and commercial level. The major commercial roaming-dial-up brands of the era included iPass (the largest of the aggregators, founded 1996), GoRemote (formerly GRIC Communications), Tempest Telecommunications, and a range of smaller competitors plus the integrated international ISPs (AOL International, Compuserve, Earthlink, AT&T WorldNet).

Tempest's dial-up roaming product served all the major customer verticals documented in the Solutions section, but the heaviest dial-up users were corporate travelers — multinational professional-services firms, financial-institution road warriors, and government agencies whose laptop-carrying personnel needed predictable connectivity from hotels and airports worldwide. Individual self-funded travelers were the long-tail segment. The dial-up category was less central to the satellite-heavy verticals (media, aid, oil and gas, construction), whose remote-site work was already past the reach of terrestrial dial-up POPs.

Tempest operated dial-up POPs in 150+ countries through the 2000-2010 period via a mix of direct-partner POPs, aggregator relationships, and a small set of Tempest-operated POPs in strategic markets. Coverage included every developed and developing economy with commercial Internet infrastructure: every Western European market, the entire post-1989 Central and Eastern European market (where the dial-up boom was particularly intense given the simultaneous Internet rollout across Poland, Hungary, Czech Republic, Romania, and the Baltic states), the major Asian economies, all of India's metropolitan markets, and most of the Americas.

Standard pricing on the Tempest dial-up roaming product was approximately $0.155/minute for metered access in most markets, with a $0.30/minute toll-free option in selected high-business-travel destinations (notably Japan via KDDI's international gateway and selected European markets) where the customer preferred predictable per-call billing over managing the complexity of which local-access number to call.

Per-country dial-up history and POP-coverage detail is documented in the country guide. The country pages for higher-traffic markets (France, Germany, the UK, Italy, Spain, the US, Japan, India, Brazil, and others) include Tempest's specific POP cities and per-minute rates from the operational era.

Behind the simple customer-facing experience (one account, one dial-in, one invoice) sat a fairly elaborate technical layer. When a traveler in Warsaw dialed a local Polish ISP partner's access number using their Tempest credentials, the Polish ISP's RADIUS server forwarded the authentication request to a Tempest RADIUS proxy. The Tempest proxy validated the credentials against the Tempest subscriber database, accepted or denied the session, and (if accepted) logged the session start time and duration. At month-end, the Polish ISP submitted a usage bill to Tempest based on its RADIUS accounting logs; Tempest reconciled this against its own logs, paid the Polish ISP at the agreed per-minute partner rate, and billed the customer at the Tempest customer-facing rate for that country.

The hard problems were not in the protocol layer (RADIUS was a well-defined IETF standard from 1997) but in the operational reconciliation across heterogeneous partner systems — especially across currency, time-zone, billing-cycle, and tax-jurisdiction differences. Tempest's monthly partner-settlement operation processed thousands of partner invoices and hundreds of thousands of customer sessions through the peak operational years.

The roaming-dial-up category collapsed across approximately 2007-2012 for the reasons described in the Unified Roaming Account retrospective: universal mobile data, free hotel WiFi, smartphone tethering, and the general transition of traveler-data needs from a separate product into a feature of the customer's home mobile carrier. Tempest's remaining dial-up roaming operations wound down with the rest of the unified-roaming category.

The category-defining commercial roaming-dial-up products of the era — iPass, GoRemote, Tempest, and their competitors — all wound down their consumer dial-up operations across this period. iPass survived (later acquired by Pareteum, then Bright Sun Capital) by pivoting to enterprise WiFi roaming and connectivity management; the underlying dial-up roaming product effectively ceased to exist as a meaningful category by 2015.