Why We Should Rebuild the Internet from the Ground Up
Les Echos’ Solveig Godeluck explains how the Internet was not originally conceived for 2.4 billion users and why it suffers from a size problem that pushes some industry pundits to lobby for a redesign or even a complete overhaul.
Videos that take hours to load, companies that keep fending off hack attempts, or email boxes invaded by spam. Users that are too anonymous for the government to track, a network that is too US-focused in the eyes of Beijing, etc. For all these reasons, and many others, researchers, businesses and governments today would change the Internet.
Some companies have already built ??an “overlay” on top of the Internet for their own purposes, such as Skype, which has quality requirements for voice and video communications, or Akamai, which proposes to speed up the flow of its customers via its own network. In addition, several research teams are working on the subject. In the United States, it’s the FIND National Science Foundation or the GENI or RINA project. In Europe it is the FIRE initiative, funded by Brussels. China and Japan are not far behind either.
Interest levels are certainly contradictory, but one thing is clear to all: a network designed to connect a few hundred scientists does not have the same requirements as a vortex sucking in more than 2.4 billion Internet users, especially when those users connect from all types of devices, do business online or handle assisted surgical procedures from remote locations.
Some researchers want to downright wipe the slate clean. John Day, who heads the RINA project in Boston, is one of the proponents of the tabula rasa. They want to replace the TCP/IP protocol created over thirty years ago by the founding fathers Vint Cerf and Bob Kahn.
Day, who admires the pioneering work of French scientist Louis Pouzin, would like to revive the diversity of the Internet’s early days. “The Internet is no longer a network of networks, but one big network, he laments. While, in the 80s, many companies created private networks, the U.S. government has offered its TCP/IP protocol to everyone, and frozen initiatives that would have lead to the development of a slower growth – but more secure – Internet.
A Design Flaw
John Day would like to fix a design flaw: on the Internet, IP addresses confuse identity and location. In other words, we give addresses to machines and not to real people. What may seem like a detail isn’t. For example, it makes the network less “redundant” – while the Internet has been created specifically to provide this type of redundancy. The IP address problem results in all sorts of constraints: allocation of network resources, mobility management…
“With a clean sweep,we could identify content or services instead of machines. It would be a complete paradigm shift,” says Kave Salamatian, professor of computer science at the University of Savoie. This would be in the best interest of companies like Google. While today they are forced to negotiate with telecom operators in order to locate content servers in their facilities, tomorrow they could claim their independence. ISPs on the other side, are obviously less enthusiastic.
“Today, with the Internet, everything changes: the upper layer applications, the infrastructure, the technology with 4G or optical fiber. Only the IP protocol has not changed, “said Kave Salamatian. However, he points out, you can also continue to put patches over the IP protocol. Stéphane Bortzmeyer, R&D engineer at AFNIC, believes, however, that the patches will prevail, as supporters of the clean slate underestimate the weight of the existing ecosystem: “We already have so much trouble with the deployment of IPv6 [the solution to the shortage of IP addresses with the original Internet] that I do not see how you could build what would not amount to a new Internet, but an entirely other network!”