Summary

  • Dai Davies helped create DANTE in 1993 as an NREN-owned, not-for-profit body able to procure, supervise and operate successive pan-European research backbones; the missing innovation was an institutional buyer, not just another protocol.
  • DANTE separated community coordination from commercial delivery. TERENA could remain a forum for technical collaboration while DANTE aggregated requirements, ran tenders, accepted networks and managed suppliers on behalf of national research and education networks.
  • Telecommunications liberalisation widened the supplier market, but it did not act alone. A contestable market became useful because DANTE had accumulated the specifications, purchasing power and operational competence needed to exploit it.

A network with no natural customer

An international backbone is easy to draw as a line across a map. It is harder to buy. A carrier sells circuits under national rules; a university wants a reliable service; a national research and education network represents many universities; a European programme may supply part of the money. None of those facts automatically produces a party able to define the whole product and hold a multinational set of suppliers accountable for it.

This was the institutional gap facing European research networking in the early 1990s. The community had technical forums, national networks and public programmes. What it still needed was a repeat customer with enough legitimacy to aggregate their demand and enough practical authority to convert it into a working service. DANTE—Delivery of Advanced Network Technology to Europe—was created in 1993 to occupy that narrow but powerful position.

Its archive describes a modest beginning: eight staff in Cambridge, about €3 million in annual turnover, and responsibility for supervising and developing Europe’s principal international research backbone. The form mattered. DANTE was a limited-liability, not-for-profit company owned by European NRENs. It was not a carrier imposed above them, nor merely another meeting. It was their instrument for acting in a market where each acting alone would have been a weaker buyer.

Davies before DANTE

Dai Davies arrived at that problem from both engineering and organisation. The Internet Hall of Fame records that he studied engineering at Cambridge, began work as a telecommunications development engineer in 1971, later studied computer science and participated in early British packet-switching development. He went on to direct COSINE, a pan-European programme intended to advance Open Systems Interconnection through pilot implementations and gateways to proprietary systems.

That history can be misread as a tale in which one protocol family defeated another. The more revealing episode is EuropaNet. In 1991, according to the Hall of Fame, Davies introduced Internet technology into a backbone originally planned as an X.25 network. The institution’s purpose survived a change in technical direction. Europe did not need to make its delivery machinery hostage to one protocol forecast.

The Hall of Fame credits COSINE with creating DANTE. DANTE’s own anniversary material names Dai Davies and Howard Davies as founding members, and records Dai as general manager from 1993 until 2011. The useful claim is not that he personally built every circuit or authored every rule. It is that his technical background and commercial role met at the point where a network programme had to become a durable operating company.

A forum and an operator are different things

European research networking already had RARE, founded in 1986, and later TERENA. A GÉANT remembrance by Fernando Liello says a detached structure became necessary to create and manage a large international backbone. In his recollection, Howard Davies was the diplomat and listener, while Dai was the more commercial manager and skilled salesman. That is a colleague’s institutional memory, not a constitutional document, but it identifies the practical division of labour.

DANTE’s own archive describes TERENA as a separate but complementary sister organisation. TERENA gave NRENs a forum: task forces, knowledge exchange, shared-service development, information and events. DANTE planned, procured, built and operated networks and coordinated delivery programmes. The difference was not between talking and doing in any dismissive sense. A community requires somewhere to deliberate; a contract requires a party that can sign, inspect, accept and, if necessary, contest it.

Separating those roles protected both. A forum could represent a plural technical community without carrying every operational liability. An operator could make time-bound commercial decisions without pretending that supplier management was the same as community consensus. DANTE’s legitimacy came from remaining an NREN-owned tool while being specialised enough to negotiate on their behalf.

Procurement as network architecture

DANTE’s strategy documents make the supposedly administrative work explicit. They describe a need for an efficient and cost-effective vehicle that could deliver pan-European projects on time, within budget and with high reliability. The organisation listed procurement, project and programme management, infrastructure development, feasibility assessment, service provision and customer support among its core competencies.

Those activities shaped the network as surely as routers did. A specification decides which cities and national networks count, how capacity can grow, where failure responsibility sits and what evidence constitutes acceptable performance. A tender decides whether suppliers compete for a coherent European service or preserve a patchwork of national bargains. Acceptance testing turns promises into an operational threshold. Migration planning determines whether switching suppliers is a credible option or merely a clause on paper.

A May 1994 Internet Monthly Report captured DANTE in this commissioning role. The organisation had received the EuroCAIRN contract to specify requirements and plan a 34–155 Mbps European research infrastructure. Davies called it an opportunity to streamline pan-European networking. The important verb was not invent but specify: expertise had been concentrated into a buyer able to tell the market what the community required.

The ladder was organisational before it was optical

DANTE’s retrospective timeline records EuropaNet2 connecting ten countries in 1995, TEN-34 serving 13 countries at up to 34 Mbps in 1997, and TEN-155 providing access capacities of 155 Mbps across 15 countries in 1998. GÉANT was built in less than a year and, by the archive’s account, served 27 European NRENs. The 2001 annual report documents a broader generation ladder from IXI at 64 Kbit/s through EuropaNet, TEN-34 and TEN-155 to GÉANT at 10 Gbit/s.

These figures are useful but dangerous. They are generation-level markers, not proof that every institution received the advertised maximum, and bandwidth alone does not measure research value. The institutional continuity is more telling. Each network generated supplier knowledge, operating evidence and a basis for the next procurement. The buyer learned. It could revise requirements, compare bids, stage migration and return to the market with a credible record.

The 2001 report shows that learning in practice. DANTE completed procurement and contracted the elements needed for GÉANT in the first half of the year, then moved through delivery, acceptance and implementation. TEN-155 stayed in service while the new infrastructure was rolled out; GÉANT became operational from 1 December. A capacity graph cannot show the risk contained by that overlap. An operator can.

Competition needed a customer

European telecommunications liberalisation changed the feasible market. European Union measures required most remaining barriers to telecommunications services and infrastructure to be removed by 1 January 1998, with transitional arrangements in some states. An EU notice stressed that access agreements were central to making liberalisation real. DANTE’s own history links the opening of formerly closed markets to falling prices and greater capacity.

The Internet Hall of Fame says Davies successfully lobbied for liberalisation. That contribution should be neither erased nor exaggerated. European law emerged from governments, EU institutions, regulators, market entrants and many advocates. Davies did not liberalise a continent alone. His strategic significance lay in articulating what research networks needed from reform and then leading an organisation capable of using the resulting choice.

Formal competition does not automatically produce effective purchasing power. A fragmented set of buyers can face nominally multiple carriers and still receive incomparable offers, incompatible terms or weak cross-border guarantees. DANTE pooled demand and learned how to procure. Liberalisation enlarged the menu; the institution made the menu actionable.

The hidden asset Europe kept

Davies led DANTE through successive generations of pan-European networks until 2011, and the model later informed collaborations in other regions. The visible inheritance is faster connectivity. The less visible one is buyer competence: people who know how to translate scientific demand into service definitions, run complex tenders, supervise deployment and preserve the ability to change suppliers.

That competence is easy to undervalue because it appears as overhead when the network works. Yet a consortium without it remains dependent on whoever already owns the lines and understands the contracts. DANTE made the research community legible to suppliers while also making suppliers answerable to the community. It converted public purpose into commercial agency without turning the network into a conventional profit-seeking carrier.

Dai Davies’s place in Internet history therefore sits between protocol and policy. He helped build the buyer Europe did not yet have. Once that buyer existed, faster networks became not a sequence of exceptional projects but a repeatable institutional practice.

Sources

  1. Internet Hall of Fame, “Dai Davies”
  2. Internet Hall of Fame, “Preserving the Internet’s Freewheeling Spirit”
  3. DANTE, “20 Years of DANTE”
  4. DANTE, “20 Years of Networking Excellence”
  5. DANTE, “About Us”
  6. DANTE Strategy 2012–2015
  7. DANTE, “TERENA — networking the networkers”
  8. DANTE, 2001 Report and Accounts
  9. GÉANT, “Howard Davies: a true British gentleman at the service of research networks”
  10. IETF, Internet Monthly Report, May 1994
  11. EUR-Lex, Commission Directive 96/19/EC
  12. EUR-Lex, notice on telecommunications access agreements