140 W.P. Worzel
Fig. 2 Aggregated data used for global models
In these scenarios, what does Fog Computing look like? Sensors may be buried
or even plowed into fields and the data can be collected during the growing
season whenever farm machinery or drones pass over the field to collect the
data. Beyond simple collection of data, each sensor, while not connected to a
central data repository such as the farm center, may be connected wirelessly to
other sensors. Since modern sensors have a great deal of processing capacity, but
potentially limited storage capacity (particularly when data is collected 24-7), by
collaboratively processing data using their combined processing capabilities, they
can “compress” the data into a more concise form. In this context, one may consider
data processing as the ultimate compression algorithm.
Similar examples may be found in weather prediction, traffic models, city
infrastructure, the Smart Grid and many manufacturing industries.
2.2 Fog Lifter Platform
In order for Fog Lifter to accomplish this goal, it must meet the following criteria:
- Organize distributed computing of local data;
- Support intermittent connectivity;
- Make locally processed results available globally;
- Integrate local results with the Cloud (remote data centers) or other fog localities
(local clusters of computational resources).