A Research Perspective on Fog Computing

State-of-the-art applications are typically deployed on top of cloud services which offer the illusion of infinite resources, elastic scalability, and a simple pay-per-use billing model. While this is very convenient for developers, it also comes with relatively high access latency for end users. Future application domains such as the Internet of Things, autonomous driving, or future 5G mobile apps, however, require low latency access which is typically achieved by moving computation towards the edge of the network. This natural extension of the cloud towards the edge is typically referred to as Fog Computing and has lately found a lot of attention. However, Fog Computing as a deployment platform has not yet found widespread adoption; this, we believe, could be helped through a consistent use of the service-oriented computing paradigm for fog infrastructure services. Based on this motivation, this paper describes the concept of Fog Computing in detail, discusses the main obstacles for Fog Computing adoption, and derives open research challenges.