The Cognitive Net Is Coming

Newswise — By 2020, according to forecasts from Cisco Systems, the global Internet will consist of 50 billion connected tags, televisions, cars, kitchen appliances, surveillance cameras, smartphones, utility meters, and whatnot. This is the Internet of Things, and what an idyllic concept it is. But here’s the harsh reality: Without a radical overhaul to its underpinnings, such a massive, variable network will likely create more problems than it proposes to solve. The reason? Today’s Internet just isn’t equipped to manage the kind of traffic that billions more nodes and diverse applications will surely bring. In fact, it’s already struggling to cope with the data being generated by evermore-popular online activities, including video streaming, voice conferencing, and social gaming. Major Internet service providers around the world are now reporting global latencies greater than 120 milliseconds, which is about as much as a voice over Internet Protocol connection can handle. Just imagine how slowly traffic would move if console gamers and cable television watchers, who now consume hundreds of exabytes of data off-line, suddenly migrated to cloud-based services. The problem is not simply one of volume. Network operators will always be able to add capacity by transmitting data more efficiently and by rolling out more cables and cellular base stations. But this approach is increasingly costly and ultimately unscalable,because the real trouble lies with the technology at the heart of the Internet: its routing architecture. It’s time we gave the Internet some smarts, not simply by making incremental improvements but by developing an entirely new way to transport data. And engineers are turning to nature for inspiration. Millions of years of evolution have resulted in biological networks that have devised ingenious solutions to the hardest network problems, such as protecting against infectious agents and adapting to failures and changes. In particular, the human brain and body are excellent models for building better data networks. The challenge, of course, is in figuring out how to mimic them.