Fog computing is an emerging technology to address computing and networking bottlenecks in large scale deployment of IoT applications. It is a promising complementary computing paradigm to cloud computing where computational, networking, storage and acceleration elements are deployed at the edge and network layers in a multi-tier, distributed and possibly cooperative manner. These elements may be virtualized computing functions placed at edge devices or network elements on demand, realizing the ''computing everywhere'' concept. To put the current research in perspective, this paper provides an inclusive taxonomy for architectural, algorithmic and technologic aspects of fog computing. The computing paradigms and their architectural distinctions, including cloud, edge, mobile edge and fog computing are subsequently reviewed. Practical deployment of fog computing includes a number of different aspects such as system design, application design, software implementation, security, computing resource management and networking. A comprehensive survey of all these aspects from the architectural point of view is covered. Current reference architectures and major application-specific architectures describing their salient features and distinctions in the context of fog computing are explored. Base architectures for application, software, security, computing resource management and networking are presented and are evaluated using a proposed maturity model. INDEX TERMS Cloud Computing, edge computing, fog computing, Internet of Things (IoT), advanced internet architecture.

No abstract

No abstract

No abstract

No abstract