We were discussing the other day in class (COM 250H) the Net neutralityprinciple, which if violated would destroy the great philosophical underpinning of the Internet: its edge-to-edge openness and simplicity. The same The Economist, which attacked the issue with great gusto in an article mentioned here previously, publishes an eye-opening material about the future of the Internet. The conclusion is that the e-2-e principle needs an update. It might not need be discarded, but it might need to be made into something a lot more flexbile (and complicated, thus error prone, would add some). In any event, the solution are “metanets,” which are networks that can use any protocol you throw at them
The internet itself stays resolutely dumb, and blindly passes packets of data between devices on the edges of the network, which can then be upgraded to do new things. This non-discriminatory approach is known as the “end-to-end principleâ€, and it is one of the most cherished aspects of the internet’s design. The end-to-end principle has promoted innovation at the edges of the network, but it has necessarily prevented innovation at its core. One possible remedy, which would allow diversity and innovation to flourish within the core of the internet, is the concept of “active networks†or “metanetsâ€.
The idea is that today’s routers, the boxes that direct internet traffic, would be replaced with more flexible devices, able to learn new communications protocols when needed. Devices at the edge of the network could then dynamically reprogram all the routers along the network path between them to use whatever new protocol they wanted. The routers would be able to partition themselves internally, so that other users of the network would not be affected. This approach has the advantage that new protocols would then emerge, tailored to applications such as video streaming, file sharing, and new things that have yet to be imagined. The drawback, however, is that routers are able to handle traffic quickly because they are single-minded. Adding complexity to the internet’s core infrastructure might end up increasing versatility at the expense of performance—a trade-off not all users would be prepared to make. [More subscription required…]