As the actions of the former Egyptian government show, it is all to easy for governments to simply shut down communications in and out of a country. Even following this, we see many western governments putting into motion plans to "secure" the internet against external (and, naturally, internal) threats. These take the form of internet kill switches in the US and government-backed eavesdropping in Sweden. Then there is the greed and avarice of the telecom companies themselves. The oligarchy charges outrageous prices, secure in the fact that no-one else can enter the business because it's impossible to buy the land, infrastructure, and right-of-way needed to set up a new telecommunications system.
For these reasons, it's obvious that a new system is needed. We need a network that is secure, free, and, well, free in the other sense too.
I believe that it is possible to make this a reality. This is Wireless Mesh Networking.
WMN works by using nodes which communicate to one another to send information along the network, using wifi or similar signals. There is no central point; all participants are equal peers, and your ability to connect to the network is limited only by the hardware you invest in. Security is paradoxically quite high; every packet is entirely anonymous, knowing only it's destination, and encrypted to the point of impenetrability.
The network is used to run a peer-to-peer http service. You can create a website and publish it, and it will be propagated around the network. There are technical difficulties in this, particularly in sharing updates and searching for content, but these are not impossible to overcome.
There are drawbacks to this network:
First; ping is severely increased at long distance, as not only must the signal travel each hop from node to node, but it must also pass through each node's routing system. However, pings between close points, such as neighboring houses, may be somewhat faster, as it wouldn't need to travel to a third party and back. That aside, it is an unavoidable fact that this would ping slower.
Second; bandwidth could be low, depending on the number of nearby nodes and the available paths from A to B. Good pathing code could actually give quite good bandwidth by sending packets along several parallel routes. This issue goes away as the mesh gets denser.
Third; long, empty regions without good places for nodes are serious barriers. We could build a network stretching from New York to Miami, but getting from London to France would be nearly impossible. Luckily, we are not restricted to purely wifi connections alone. It could be possible to connect over internet packet radio between isolated sections to create truly global networks that are completely independent and free. It would also be possible to use existing internet connections between separated networks. Failing that, we could simply rely on geographically isolated networks that are only updated intermittently.
My vision is simple: You can take an off-the-shelf router, install some new firmware on it and get a special browser on your computer, and find yourself connected to the new internet- for free, free from snooping and blackouts, and free to make full use of it without paying anyone.