The problem is that outer space is likely to be as hostile to the networking protocols that underpin the Internet as it is to air breathing humanoids. For a start TCP/IP, the common language spoken by every device on the network, is a very chatty protocol. A constant buzz of greetings, acknowledgments and farewells travel between computers as they locate each other, exchange information and then disconnect. In space however such chattiness is a bad way of conversing. Radio signals take a second to reach the moon, and several minutes to reach Mars. So a new, terser protocol will be needed, to save both time and energy.

Another difficulty is that TCP/IP was designed to work over networks in which transmission errors are rare, but congestion is common. On earth when one device sends a packet of data to another and fails to receive an acknowledgment, it assumes that the link has become congested. It therefore reduces the rate at which it sends subsequent packets, only ramping that rate up again when the congestion has eased. In space, however, different rules apply. If a packet of data sent to a distant spacecraft in a hostile environment fails to get through, the cause is more likely to be a transmission error, in which case the correct response is to retransmit the packet as quickly as possible, rather than assuming that the link is congested and backing off.

File transfer is another area where new approaches will be required. At the moment sending a file (an image, for example) from one computer to another using the Internet’s File Transfer Protocol (FTP) involves establishing a connection between the source and destination machines, and then passing the file across in chunks. But if a rover on the Martian surface wants to send a file back to earth, this is an inefficient way of doing it. It would make more sense for the rover to hand the whole file over to a lander, which could pass it in chunks to orbiting relay satellites for transmission home.

Consideration of such matters sounds whimsical. But similar problems arise, in less drastic form, with wireless Internet devices on earth. Handheld computers and wireless netphones would also benefit from a less chatty protocol, more efficient use of their limited battery power, the ability to cope in noisy environments, and an easier way to send files while on the move.

So rather than reinventing the wheel, the scientists working on IPN a consortium that includes researchers from America’s space agency, NASA, the Defence Advanced Research Projects Agency, the British National Space Centre and Britain’s Defence Evaluation and Research Agency, hope to collaborate with researchers in the terrestrial telecoms industry to establish new standards. The IPN working group has already drawn up a list of problems that need to be solved.

So far, the plans include the development of File Delivery Protocol (FDP), a modified form of FTP, and a new idea based around bundles of data in which multiple packets, requests, files and messages can be sent in one go.

The idea, eventually, is that separate Internets should exist on Earth, the moon, Mars and so on, connected by gateways that communicate over an interplanetary backbone using new spacefaring protocols. Probes landing on asteroids and comets would also connect to the IPN. The fact that different bodies in the solar system change their relative positions as they orbit the sun will complicate routing.

There is also the question of names and addresses. The IPN plans call for an extension to the domain naming system to cover different planets and solar systems.

For example, websites on Earth would become www.website.com.earth.sol