Fog and Fogonomics

4Pricing Tradeoffs for Data Analytics in Fog–Cloud Scenarios

Yichen Ruan¹, Liang Zheng², Maria Gorlatova², Mung Chiang³, and Carlee Joe‐Wong¹

¹Department of Electrical and Computer Engineering, Carnegie Mellon University, Moffett Field, CA, USA

²Department of Electrical Engineering, Princeton University, Princeton, NJ, USA

³Department of Electrical and Computer Engineering, Purdue University, West Lafayette, IN, USA

4.1 Introduction: Economics and Fog Computing

The proliferation of devices (sometimes called “things”) and enriched applications has created an expectation of near‐ubiquitous connectivity among users around the world. Until recently, this activity has largely been enabled by cloud computing, which allows users to virtually share computing resources on a large scale. Applications can then easily access these resources without spending significant amounts of money and effort on setting up and maintaining their own host servers. Cloud resources can also be easily scaled up, allowing applications to rapidly grow their user base. For instance, applications can use cloud servers to collect, store, and analyze data about their users; or to host large files like videos to be streamed. Applications and devices ranging from remote‐access home surveillance cameras to health monitoring smartwatches rely on the cloud to serve their users today.

The traditional cloud paradigm, in which applications utilize servers in a remote data center, may not meet new performance requirements ...