As this feature in The Economist explains, "The bottlenecks -- whether at the DNS translators, the routing computers or the host's own servers -- stem largely from the way the mix of Internet traffic has changed faster than the infrastructure used to carry it."
In short, the Web is like a war zone, where pockmarked roads and aging checkpoints force data to take detour after detour in its travels. Each detour adds time, and microseconds eventually add up to seconds. Lighting up a parallel fiber network could makes these detours irrelevant, and add speed.
Google needs the Web to be faster than it is, and it can't wait for telcos and ISPs to act. On Wednesday, the search king announced a plan to offer fiber-to-the-home connections to as many as 500,000 people in a handful of yet-to-be-named U.S. communities.
Is this really necessary?
Yes. Real-time delivery of data is essential to Google, because every one of its services is delivered, rather than installed. The Web makes Google possible; the Real-Time Web makes Google the next Microsoft.
How? Imagine if Web-delivered services loaded as fast as installed software does today. Now imagine if those same services were as functional as their desktop peers, yet sold for a fraction of the cost. That's the future salesforce.com
It's less clear whether Google's next step will be to take this from experiment to full-blown project. That cause could lead the company to spend the $8 billion-$10 billion (after debt assumption) needed to acquire Level 3's
What do you think? Should Google roll out its own national broadband plan? Should it spend the money to acquire Level 3? Or should this "experiment" remain just that? Make your voice heard using the comments box below.