Abstract
AbstractRecent advances in mobile connectivity as well as increased computational power and storage in sensor devices have given rise to a new family of software architectures with challenges for data and communication paths as well as architectural reconfigurability at runtime. Established in 2012, Fog Computing describes one of these software architectures. It lacks a commonly accepted definition, which manifests itself in the missing support for mobile applications as well as dynamically changing runtime configurations. The dissertation “Dynamically Scalable Fog Architectures” provides a framework that formalizes Fog Computing and adds support for dynamic and scalable Fog Architectures.The framework called xFog (Extension for Fog Computing) models Fog Architectures based on set theory and graphs. It consists of three parts: xFogCore, xFogPlus, and xFogStar. xFogCore establishes the set theoretical foundations. xFogPlus enables dynamic and scalable Fog Architectures to dynamically add new components or layers. Additionally, xFogPlus provides a View concept which allows stakeholders to focus on different levels of abstraction.These formalizations establish the foundation for new concepts in the area of Fog Computing. One such concept, xFogStar, provides a workflow to find the best service configuration based on quality of service parameters.The xFog framework has been applied in eight case studies to investigate the applicability of dynamic Fog Components, scalable Fog Architectures, and the service provider selection at runtime. The case studies, covering different application domains—ranging from smart environments, health, and metrology to gaming—successfully demonstrated the feasibility of the formalizations provided by xFog, the dynamic change of Fog Architectures by adding new components and layers at runtime, as well as the applicability of a workflow to establish the best service configuration.
Highlights
With the constant increase of computational power and available storage, mobile devices get more and more involved in distributed systems, which are “collections of independent computers that appear to be one single system to users” [1]
Mobile cloud computing was introduced to bridge this gap and combines mobile computing with cloud computing to leverage the computational power of the cloud for mobile devices [4, 5]
The main goal of the dissertation “Dynamically Scalable Fog Architectures” [10] was to create a framework that establishes a formalization for Fog Computing and integrates support for mobile applications and dynamic reconfigurability of Fog Architectures
Summary
With the constant increase of computational power and available storage, mobile devices get more and more involved in distributed systems, which are “collections of independent computers that appear to be one single system to users” [1]. Clouds are usually distant from the mobile devices and using them creates high latencies, which are insufficient for real-time applications such as augmented reality To address this issue, concepts such as Cloudlets, Edge Computing, and Fog Computing emerged. There is no commonly accepted definition of what Fog Computing or a Fog Node is, and the difference to similar concepts such as Edge Computing is not clearly defined [9] To address these misunderstandings, the dissertation “Dynamically Scalable Fog Architectures” [10] follows the intent of the paper “Fog horizons—a theoretical concept to enable dynamic fog architectures” [17] to create a formalized definition for Fog Computing based on software architectures and set theory. Technical Research Goal 3: Define an extension to the foundations of Fog Computing that supports the process of adding and removing layers (Scalability) These goals should be addressed with xFogPlus, one part of xFog that relies on the foundations introduced by xFogCore and formalizes dynamics and scalability in Fog Architectures. XFogStar extends xFog with a workflow that supports service consumers with the selection of service providers
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
