Cambridge boffins develop mobile technology without cells.

A new programming language has been developed to try to reduce the complexity of building applications for ad hoc peer-to-peer mobile device networks. The new D3N (Data-Driven Declarative Networking) language devloped by a team from Cambridge University offers a way to communicate without some of the geographical constraints.

Research team leader, Jon Crowcroft  released a paper paper describing the new technology. "We envision an emergence of a new type of communications based on physical proximity, where people encounter each other and devices directly communicate within their range," the paper said. "We introduce declarative networking for pocket-switched networks called D3N, which allows applications to construct a protocol description consisting of a reactive behaviour of a distributed node."

Such peer-to-peer mobile device networks, called "pocket-switched networks," could allow communications between wireless devices without the need for conventional cellular networks, thereby eliminating the need for cell towers and base stations, researchers said in a recent issue of MIT's Technology Review.

For example, an ad hoc network would mean that after a hurricane knocks down cell towers, as happened when Hurricane Katrina hit near New Orleans in 2005, people using wireless devices could theoretically connect with one another over pocket-switched networks for access to help. Or, instead of going to the Internet for information on good places to stay or eat in an unfamiliar town, an ad hoc mobile network might help a visitor get the information from local residents or businesses.

Peer-to-peer networking is a fairly commonplace idea in wired networks, but linking wireless smartphones and other handheld computers through pocket-switched networks would require using Bluetooth or another short-range wireless network.

Crowcroft told Technology Review that if D3N gains popularity, it would usher in a "whole slew of applications." Part of the challenge of building an ad hoc wireless network would be dealing with delays in connections, partly because individual mobile devices might be only occasionally connected. Still, they would need to be able to forward messages when reconnected. Users would subscribe to be part of an ad hoc network.

D3N is based on the F# programming language, a Microsoft initiative. The paper says that future research will include validating the D3N language on a compiler deployed on actual wireless devices.