What is the most effective way to take 1 PetaBytes of data from San Francisco to London? Put them on hard drives and ship the drives on a jumbo jet. They’ll get there sooner than if you were to use the internet (at a connection speed of 1Gbps transferring it will take about 3 months to transfer all the data, on the other hand you can put 250 hard drives 4TB each on a 747 and have the transfer completed in 10 hours).
This extreme case is to underly that there may be different ways to transport data from A to B than using the Internet.
Imagine you have sensors spread all around a city. Connecting each of these sensors to the telecommunication network may be a challenge in terms of cost and of power. What if one could piggy back on vehicles for transporting the bits? In a city there are plenty of vehicles moving around in a randomic way (if you are looking at them all). Hence there is a quite high probability that a vehicle can be in the neiborogh of a sensors every few hours (even every few minutes depending on the sensor location). Why not having the sensor transmit its data to the vehicle and let the vehicle transport them to the point where they can be uploaded on the internet to go to their final destination? The vehicle may actually transfer the data to other vehicles, let’s say a hundred of them, thus increasing to certainty the probability that one of them will actually get close enough to a internet access point.
These types of networks are known as sneakernets, since they rely on “sneakers” of some sort to carry the data.
Moore’s Law is making these kind of nets feasible by decreasing the power needed for communications (here we are talking of a few meters…. so the radio signal can be really weak) and by slashing the cost of storage so that having some storage space to transport data around is basically cost-free.
Sneakernets can be an interesting solution for Smart Cities since in a city you have plenty of things moving around (not just private vehicles: you got public transportation, bikes, taxis, even birds -like upgraded pidgeons- can be considered, and goods being transported).
The paradigm is not that different from the communication paradigm used for centuries in the past. At that time communications happened by word of mouth: one person telling a story to another and the message getting spread as people moved around. It might seem not a very effective way of communications but it worked.
Notice that there are several “instantiations” of this paradigm.
As an example, rather than clogging the wireless space with data transport why not have data distributed throughout the city to some data beacons, located in main shopping centers, squares, bus stops… These beacons can be updated every time something new happens and cell phones passing nearby can capture the data updating their “news and fact storage”.
There is more. As cell phones storage capacity increases, thanks to Moore’s Law, into the TB range there will be plenty of “sneakers” to take data around. Cell phones are ideal to act as data repository, beacons and create geographic and social related sneakernets. I am looking forward to applications similar to BitTorrent applied to the creation of sneakernets based on smartphones. Clearly a Municipality would have a hard time to “force” citizens to release part of their smart phones storage to become sneakers. But this is not the right approach. One should go into the opt-in strategy coupled with a service strategy. You can benefit from certain services if at the same time you are willing to share part of your phone processing and storage capacity. More than that. A service may “imply/require” a sharing of data to be useful. We have plenty of evidence that people, citizens, are willing to share. Let’s leverage on that.